Integrated Pest Management: Concepts, Tactics, Strategies and Case Studies

  • 82 422 7
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

Integrated Pest Management: Concepts, Tactics, Strategies and Case Studies

Integrated Pest Management Concepts, Tactics, Strategies and Case Studies Integrated Pest Management (IPM) is an effecti

893 432 14MB

Pages 549 Page size 561.75 x 737.25 pts Year 2011

Report DMCA / Copyright


Recommend Papers

File loading please wait...
Citation preview

Integrated Pest Management Concepts, Tactics, Strategies and Case Studies Integrated Pest Management (IPM) is an effective and environmentally sensitive approach to pest management. It uses natural predators, pest-resistant plants and other methods to preserve a healthy environment in an effort to decrease reliance on harmful pesticides. Featuring 40 chapters written by leading experts, this textbook covers a broad and comprehensive range of topics in Integrated Pest Management, focused primarily on theory and concepts. It is complemented by two award-winning websites, which are regularly updated and emphasize specific IPM tactics, their application, and IPM case studies: Radcliffe’s IPM World Textbook – http://ipmworld. VegEdge – The two products are fully cross-referenced and form a unique and highly valuable resource. Written

with an international audience in mind, this text is suitable for advanced undergraduate and graduate courses on Integrated Pest Management, Insect or Arthropod Pest Management. It is also a valuable resource for researchers, extension specialists and IPM practitioners worldwide. Edward B. Radcliffe is a Professor in the Department of Entomology, University of Minnesota, St. Paul, where he has taught IPM since 1966. William D. Hutchison is Professor and Extension Entomologist in the Department of Entomology, University of Minnesota, St. Paul. Rafael E. Cancelado is an independent crop consultant working with vegetable growers in Lara Region, Venezuela.

Integrated Pest Management Concepts, Tactics, Strategies and Case Studies Edited by

Edward B. Radcliffe University of Minnesota, St. Paul

William D. Hutchison University of Minnesota, St. Paul

Rafael E. Cancelado Venezuela

cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, S˜ ao Paulo, Delhi Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York Information on this title:  c Cambridge University Press 2009

This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2009 Printed in the United Kingdom at the University Press, Cambridge A catalog record for this publication is available from the British Library Library of Congress Cataloging in Publication data Integrated pest management : concepts, tactics, strategies & case studies / edited by Edward B. Radcliffe, William D. Hutchison, Rafael E. Cancelado. p. cm. Includes index. ISBN 978-0-521-87595-0 (hardback) 1. Pests – Integrated control. I. Radcliffe, Edward B., 1936– II. Hutchison, William D. III. Cancelado, Rafael E. IV. Title. SB950.I4572 2008 632 .9 – dc22 2008042307 ISBN 978-0-521-87595-0 hardback ISBN 978-0-521-69931-0 paperback

Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.


This Cambridge University Press IPM textbook is dedicated to the memory of Robert (Bob) J. O’Neil (1955–2008), who died on February 6 following an almost year-long battle against bladder cancer. Bob worked on the chapter he co-authored with John Obrycki for this book throughout his illness. During his professional career at Purdue University, Bob established an internationally recognized research, teaching and outreach program in biological control. His research and outreach accomplishments serve as a model for how undertaking fundamental research on ecologically important questions can lead to solutions of environmental problems. One of Bob’s greatest joys and lasting contributions will be his work with students from the Pan American School of Agriculture (Zamorano) in Honduras. His efforts resulted in opportunities for hundreds of Latin American students to work, study and pursue advanced degrees at Purdue and several universities in the United States and Europe.

Bob was the creative spark who acted as the catalyst to bring together individuals from several universities in the Midwestern USA to work together on biological control projects. His idea to create collaborative summer workshops for graduate students through the Midwest Biological Control Institute is now in its 18th year. His recognition of the need to enhance the level of knowledge about biological control among extension agents resulted in several multi-state extension workshops on biological control. Bob knew that biological control could contribute to the management of pests in the Midwest and formed cooperative research teams to find new natural enemies to address these pest problems. Working together with Bob, this group of individuals was able to significantly advance implementation of biological control in the Midwestern USA. Bob’s vision, can-do attitude, humor and optimistic view of the world (he was a lifelong Boston Red Sox fan who occasionally rooted for the Cubs) will be greatly missed by his colleagues and friends.


Professor Chris Curtis, one of the world’s leading medical entomologists, died unexpectedly in May 2007. Chris was completely dedicated to the cause of useful science in the service of practical public health. In his early career, he worked on genetic methods of insect control, but most of his professional life was devoted to development of low technology methods for mosquito control. In the last two decades, Chris played a major role in developing and promoting use of insecticide-treated mosquito nets (ITN) for the prevention of malaria transmission by Anopheles mosquitoes, as this technique gradually moved to center-stage for practical malaria control. His team demonstrated that, when ITNs are used by most people in a village, there is a “mass effect” on the local mosquito population that reduces its ability to transmit malaria and gives extra protection to all, including those without nets. Chris was a tireless and highly influential campaigner

for the cause of “free nets,” against the notion that nets should be sold, or targeted only at subgroups most vulnerable to malaria. In doing so, he contributed to a significant strengthening of political will in developed countries, and thus to vastly increased donor funding for malaria control. To date, about 50 million treated nets have been given away, preventing tens of thousands of deaths due to malaria among African children. As a person, Chris was exceptionally gentle, honest and kind. As a teacher he was positively luminous: he gave lasting inspiration to countless students, and there is a worldwide community of students and colleagues (including the writer) who are linked by what Chris taught them about how to do science, how to value it, and how to enjoy it. jo lines London School of Hygiene and Tropical Medicine


List of contributors Preface Acknowledgements

Chapter 1 The IPM paradigm: concepts, strategies and tactics

page xi xvii xix


Michael E. Gray, Susan T. Ratcliffe & Marlin E. Rice

Chapter 2 Economic impacts of IPM


Scott M. Swinton & George W. Norton

Chapter 3 Economic decision rules for IPM


Leon G. Higley & Robert K. D. Peterson

Chapter 4 Decision making and economic risk in IPM


Paul D. Mitchell & William D. Hutchison

Chapter 5 IPM as applied ecology: the biological precepts


David J. Horn

Chapter 6 Population dynamics and species interactions


William E. Snyder & Anthony R. Ives

Chapter 7 Sampling for detection, estimation and IPM decision making


Roger D. Moon & L. T. (Ted) Wilson

Chapter 8 Application of aerobiology to IPM


Scott A. Isard, David A. Mortensen, Shelby J. Fleischer & Erick D. De Wolf

Chapter 9 Introduction and augmentation of biological control agents


Robert J. O’Neil & John J. Obrycki

Chapter 10 Crop diversification strategies for pest regulation in IPM systems Miguel A. Altieri, Clara I. Nicholls & Luigi Ponti




Chapter 11 Manipulation of arthropod pathogens for IPM


Stephen P. Wraight & Ann E. Hajek

Chapter 12 Integrating conservation biological control into IPM systems


Mary M. Gardiner, Anna K. Fiedler, Alejandro C. Costamagna & Douglas A. Landis

Chapter 13 Barriers to adoption of biological control agents and biological pesticides


Pamela G. Marrone

Chapter 14 Integrating pesticides with biotic and biological control for arthropod pest management


Richard A. Weinzierl

Chapter 15 Pesticide resistance management


Casey W. Hoy

Chapter 16 Assessing environmental risks of pesticides


Paul C. Jepson

Chapter 17 Assessing pesticide risks to humans: putting science into practice


Brian Hughes, Larry G. Olsen & Fred Whitford

Chapter 18 Advances in breeding for host plant resistance


C. Michael Smith

Chapter 19 Resistance management to transgenic insecticidal plants


Anthony M. Shelton & Jian-Zhou Zhao

Chapter 20 Role of biotechnology in sustainable agriculture Jarrad R. Prasifka, Richard L. Hellmich & Michael J. Weiss



Chapter 21 Use of pheromones in IPM


Thomas C. Baker

Chapter 22 Insect endocrinology and hormone-based pest control products in IPM


Daniel Doucet, Michel Cusson & Arthur Retnakaran

Chapter 23 Eradication: strategies and tactics


Michelle L. Walters, Ron Sequeira, Robert Staten, Osama El-Lissy & Nathan Moses-Gonzales

Chapter 24 Insect management with physical methods in pre- and post-harvest situations


Charles Vincent, Phyllis G. Weintraub, Guy J. Hallman & Francis Fleurat-Lessard

Chapter 25 Cotton arthropod IPM


Steven E. Naranjo & Randall G. Luttrell

Chapter 26 Citrus IPM


Richard F. Lee

Chapter 27 IPM in greenhouse vegetables and ornamentals


Joop C. van Lenteren

Chapter 28 Vector and virus IPM for seed potato production


Jeffrey A. Davis, Edward B. Radcliffe, Willem Schrage & David W. Ragsdale

Chapter 29 IPM in structural habitats


Stephen A. Kells

Chapter 30 Fire ant IPM


David H. Oi & Bastiaan M. Drees

Chapter 31 Integrated vector management for malaria


Chris F. Curtis

Chapter 32 Gypsy moth IPM Michael L. McManus & Andrew M. Liebhold





Chapter 33 IPM for invasive species


Robert C. Venette & Robert L. Koch

Chapter 34 IPM information technology


John K. VanDyk

Chapter 35 Private-sector roles in advancing IPM adoption


Thomas A. Green

Chapter 36 IPM: ideals and realities in developing countries


Stephen Morse

Chapter 37 The USA National IPM Road Map


Harold D. Coble & Eldon E. Ortman

Chapter 38 The role of assessment and evaluation in IPM implementation


Carol L. Pilcher & Edwin G. Rajotte

Chapter 39 From IPM to organic and sustainable agriculture


John Aselage & Donn T. Johnson

Chapter 40 Future of IPM: a worldwide perspective


E. A. (Short) Heinrichs, Karim M. Maredia & Subbarayalu Mohankumar Index



1 The IPM paradigm: concepts, strategies and tactics Michael E. Gray University of Illinois, Crop Sciences, Urbana–Champaign, Illinois, USA Susan T. Ratcliffe University of Illinois, Crop Sciences, Urbana–Champaign, Illinois, USA Marlin E. Rice Iowa State University, Department of Entomology, Ames, Iowa, USA 2 Economic impacts of IPM Scott M. Swinton Michigan State University, Department of Agricultural Economics, East Lansing, Michigan, USA George W. Norton Virginia Polytechnic Institute and State University, Agricultural and Applied Economics, Blacksburg, Virginia, USA

5 IPM as applied ecology: the biological precepts David J. Horn The Ohio State University, Department of Entomology, Columbus, Ohio, USA 6 Population dynamics and species interactions William E. Snyder Washington State University, Department of Entomology, Pullman, Washington, USA Anthony R. Ives University of Wisconsin, Department of Zoology, Madison, Wisconsin, USA 7 Sampling for detection, estimation and IPM decision making Roger D. Moon University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA L. T. (Ted) Wilson Texas A&M, University, Agricultural Research and Extension Center, Beaumont, Texas, USA

3 Economic decision rules for IPM Leon G. Higley University of Nebraska, Department of Entomology, Lincoln, Nebraska, USA Robert K. D. Peterson Montana State University, Department of Land Resources and Environmental Sciences, Bozeman, Montana, USA 4 Decision making and economic risk in IPM

8 Application of aerobiology to IPM Scott A. Isard The Pennsylvania State University, Department of Plant Pathology, University Park, Pennsylvania, USA David A. Mortensen The Pennsylvania State University, Department of Agronomy, University Park, Pennsylvania, USA

Paul D. Mitchell University of Wisconsin, Department of Agricultural and Applied Economics, Madison, Wisconsin, USA

Shelby J. Fleischer The Pennsylvania State University, Department of Entomology, University Park, Pennsylvania, USA

William D. Hutchison University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA

Erick D. De Wolf The Pennsylvania State University, Department of Plant Pathology, University



Park, Pennsylvania, USA, currently with Kansas State University, Department of Plant Pathology, Manhattan, Kansas, USA 9 Introduction and augmentation of biological control agents Robert J. O’Neil (deceased) Purdue University, Department of Entomology, West Lafayette, Indiana, USA John J. Obrycki University of Kentucky, Department of Entomology, Lexington, Kentucky, USA 10 Crop diversification strategies for pest regulation in IPM systems Miguel A. Altieri University of California, Department of Environmental Science, Policy and Management, Division of Insect Biology, Berkeley, California, USA Clara I. Nicholls University of California, Department of Environmental Science, Policy and Management, Division of Insect Biology, Berkeley, California, USA Luigi Ponti University of California, Department of Environmental Science, Policy and Management, Division of Insect Biology, Berkeley, California, USA 11 Manipulation of arthropod pathogens for IPM Stephen P. Wraight United States Department of Agriculture, Agricultural Research Service, Plant Protection Research Unit, USA Ann E. Hajek Cornell University, Department of Entomology, Ithaca, NY, USA

12 Integrating conservation biological control into IPM systems Mary M. Gardiner Michigan State University, Department of Entomology, East Lansing, Michigan, USA Anna K. Fiedler Michigan State University, Department of Entomology, East Lansing, Michigan, USA Alejandro C. Costamagna Michigan State University, Department of Entomology, East Lansing, Michigan, currently with University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA Douglas A. Landis Michigan State University, Department of Entomology, East Lansing, Michigan, USA 13 Barriers to adoption of biological control agents and biological pesticides Pamela G. Marrone Marrone Organic Innovations, Davis, California, USA 14 Integrating pesticides with biotic and biological control for arthropod pest management Richard A. Weinzierl University of Illinois, Crop Sciences, Urbana–Champaign, Illinois, USA 15 Pesticide resistance management Casey W. Hoy The Ohio State University, Ohio Agricultural Research and Development Center, Department of Entomology, Wooster, Ohio, USA 16 Assessing environmental risks of pesticides Paul C. Jepson Oregon State University, Environmental and Molecular Toxicology Department, Corvallis, Oregon, USA


17 Assessing pesticide risks to humans: putting science into practice Brian Hughes Dow Chemical Company, Midland, Michigan, USA Larry G. Olsen Michigan State University, Department of Entomology, East Lansing, Michigan, USA Fred Whitford Purdue University, Department of Entomology, West Lafayette, Indiana, USA 18 Advances in breeding for host plant resistance C. Michael Smith Kansas State University, Department of Entomology, Manhattan, Kansas, USA 19 Resistance management to transgenic insecticidal plants Anthony M. Shelton Cornell University, Department of Entomology, Geneva, New York, USA Jian-Zhou Zhao Cornell University, Department of Entomology, Geneva, New York, USA, currently with Pioneer Hi-Bred International, Inc., a DuPont Company, Johnson, Iowa, USA 20 Role of biotechnology in sustainable agriculture Jarrad R. Prasifka Energy Biosystems Institute, Institute for Genome Biology, University of Illinois, Urbana, Illinois, USA Richard L. Hellmich United States Department of Agriculture, Agricultural Research Service, Corn Insects and Crop Genetics Research Unit, Ames, Iowa, USA Michael J. Weiss Golden Harvest Seeds, Decorah, Iowa, USA

21 Use of pheromones in IPM Thomas C. Baker The Pennsylvania State University, Department of Entomology, University Park, Pennsylvania, USA 22 Insect endocrinology and hormone-based pest control products in IPM Daniel Doucet Canadian Forest Service, Great Lakes Forestry Centre, Sault Ste. Marie, Ontario, Canada Michel Cusson Canadian Forest Service, Laurentian Forestry Centre, Quebec City, Canada Arthur Retnakaran Emeritus, Canadian Forest Service, Great Lakes Forestry Centre, Sault Ste. Marie, Ontario, Canada 23 Eradication : strategies and tactics Michelle L. Walters United States Department of Agriculture, Animal Plant Health and Inspections Service, Center for Plant Health Science and Technology, Phoenix, Arizona, USA Ron Sequeira United States Department of Agriculture, Animal Plant Health and Inspection Service, Center for Plant Health Science and Technology, North Carolina, USA Robert Staten Emeritus, United States Department of Agriculture, Animal Plant Health and Inspection Service Center for Plant Health Science and Technology, Phoenix, Arizona, USA Osama El-Lissy United States Department of Agriculture, Animal Plant Health and Inspections Service, Invasive Species and Pest Management, Riverdale, Maryland, USA Nathan Moses-Gonzales United States Department of Agriculture, Animal Plant Health and Inspections Service




Center for Plant Health Science and Technology, Phoenix, Arizona, USA 24 Insect management with physical methods in pre- and post-harvest situations Charles Vincent Agriculture and Agri-Food Canada, Horticultural Research and Development Centre, Saint-Jean-sur-Richelieu, Quebec, Canada Phyllis G. Weintraub Agricultural Research Organization, Gilat Research Center, Israel, D. N. Negev, Israel Guy J. Hallman United States Department of Agriculture, Agricultural Research Service, Subtropical Agricultural Research Center, Weslaco, Texas, USA Francis Fleurat-Lessard INRA, Laboratory for Post-Harvest Biology and Technology, Villenave-d’Ornon, France 25 Cotton arthropod IPM Steven E. Naranjo United States Department of Agriculture, Agricultural Research Service, Arid-Land Agricultural Research Center, Maricopa, Arizona, USA Randall G. Luttrell University of Arkansas, Department of Entomology, Fayetteville, Arkansas, USA 26 Citrus IPM Richard F. Lee United States Department of Agriculture, Agricultural Research Service, Germplasm Resources Information Network, National Clonal Germplasm Repository for Citrus and Dates, Riverside, California, USA 27 IPM in greenhouse vegetables and ornamentals Joop C. van Lenteren Wageningen University, Laboratory of Entomology, Wageningen, The Netherlands

28 Vector and virus IPM for seed potato production Jeffrey A. Davis University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA, currently with Louisiana State University, Department of Entomology, Baton Rouge, Louisiana, USA Edward B. Radcliffe University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA David W. Ragsdale University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA Willem Schrage Minnesota Department of Agriculture, Potato Program, East Grand Forks, Minnesota, USA, currently with North Dakota State Seed Department, Fargo, North Dakota, USA 29 IPM in structural habitats Stephen A. Kells University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA 30 Fire ant IPM David H. Oi United States Department of Agriculture, Agricultural Research Service, Center for Medical, Agricultural, and Veterinary Entomology, Gainesville, Florida, USA Bastiaan (Bart) M. Drees Texas A&M University, Department of Entomology, College Station, Texas, USA 31 Integrated vector management for malaria Chris F. Curtis (deceased) London School of Hygiene and Tropical Medicine, London, UK 32 Gypsy moth IPM Michael L. McManus Emeritus, United States Forest Service, Northeast Research Station, Hamden, Connecticut, USA


Andrew M. Liebhold United States Forest Service, Northeastern Area, Morgantown, West Virginia, USA

Eldon E. Ortman Emeritus, Purdue University, Department of Entomology, West Lafayette, Indiana, USA

33 IPM for invasive species Robert C. Venette United States Forest Service, Northern Research Station, Biological and Environmental Influences on Forest Health and Productivity, St. Paul, Minnesota, USA Robert L. Koch University of Minnesota, Department of Entomology, St. Paul, Minnesota, USA

38 The role of assessment and evaluation in IPM implementation Carol L. Pilcher Iowa State University, Department of Entomology, Ames, Iowa, USA Edwin G. Rajotte The Pennsylvania State University, Department of Entomology, University Park, Pennsylvania, USA

34 IPM information technology John K. VanDyk Iowa State University, Department of Entomology, Ames, Iowa, USA 35 Private-sector roles in advancing IPM adoption Thomas A. Green IPM Institute of North America, Inc., Madison, Wisconsin, USA

39 From IPM to organic and sustainable agriculture John Aselage Amy’s Kitchen, Santa Rosa, California, USA Donn T. Johnson University of Arkansas, Department of Entomology, Fayetteville, Arkansas, USA 40 Future of IPM: a worldwide perspective

36 IPM: ideals and realities in developing countries Stephen Morse University of Reading, International Development Centre, Applied Development Studies, Department of Geography, Reading, UK 37 The USA National IPM Road Map Harold D. Coble United States Department of Agriculture, Office of Pest Management Policy, North Carolina State University, Raleigh, North Carolina, USA

E. A. (Short) Heinrichs Emeritus, Department of Entomology, University of Nebraska, Lincoln, Nebraska, USA Karim M. Maredia Michigan State University, Institute of International Agriculture and Department of Entomology, East Lansing, Michigan, USA Subbarayalu Mohankumar Tamil Nadu Agricultural University, Department of Plant Molecular Biology and Biotechnology, Coimbatore, India



Integrated Pest Management (IPM) has been taught in the Department of Entomology at the University of Minnesota since 1966. Over the years, we’ve used many different textbooks for this course, supplementing these with primary references and more recently with web resources. We’ve never lacked for quality information resources to use in teaching our course, especially so in recent years, but we’ve never felt satisfied that any one textbook provided the breath of coverage of all the IPM related topics we think need to be included in a university-level course. We recognized that our expectations might be unrealistic since such broad coverage could make for a book of such size and cost that it wouldn’t be appropriate to adopt as a required textbook. We attempted to overcome these challenges by developing our own online textbook, Radcliffe’s IPM World Textbook, Our concept for creating this website was that we’d solicit content from a cadre of internationally recognized experts with the goal eventually of a comprehensive online IPM resource having “chapters” covering all aspects of IPM. Our primary objectives in creating this website were to provide (1) a venue for easily maintaining and updating “state of the art” information from the world’s leading experts on all aspects of IPM, and (2) a resource economically deliverable anywhere in the world that could be freely downloaded for use by students, teachers and IPM practitioners. Since 1996, we’ve used this resource, supplemented with an electronic library of primary references and links to other IPM websites, as the textbook for our teaching of IPM. This website has achieved considerable success and recognition, but coverage of topics is still uneven, and we believe there is still need for a comprehensive, printed IPM textbook. In late 2005, Cambridge University Press Commissioning Editor Jacqueline Garget suggested that we consider submitting a proposal to the Press for the development of a printed IPM textbook. The concept we developed, and that we

believed would make for a book unique among its peers, was that the new printed textbook and our existing online textbook should be complementary and cross-referenced. Our idea was that the printed textbook would focus on theory, i.e., concepts and guiding principles, and that it would provide information of general application that would not become quickly dated, whereas information and specific examples that are more timesensitive or situation-specific would be posted online. Again, we proposed creating a multiauthored textbook with the contributed chapters following the outline of a typical IPM course. To achieve that, we invited contributors to this book to write their chapters in the style of a classroom lecture. We asked that authors emphasize those key concepts they would want to communicate were they invited to present a guest lecture on their chapter topic to an undergraduate/graduatelevel IPM class. To keep this book to a reasonable size, the chapters in this work are shorter and generally contain fewer specifics and/or examples than is typical of chapters in more traditionally organized IPM textbooks. The complementary online textbook allows us to make available supplemental material including colored illustrations, searchable lists, detailed case studies and much more, all of which being online can be conveniently updated as appropriate. The terminology “Integrated Control” entered the lexicon of economic entomology almost 50 years ago. The concepts of Integrated Control, soon renamed Integrated Pest Management, were quickly embraced by the scientific community and officially accepted as the operative pest management paradigm by most governments and international organizations. Nevertheless, pesticide use continues to grow and to be the tactic of primary reliance for most pest management practitioners. It is appropriate to ask why this is so, and why IPM has not been more fully adopted. The authors of this book have addressed many of the constraints that have slowed IPM adoption, but they also present convincing arguments that IPM



remains the most robust, ecologically sound and socially desirable approach to addressing pest control challenges. In summary, we hope that readers from many perspectives will find this book interesting and of practical value. Specifically, we trust that the book, along with the complementary IPM World Textbook website, will continue to be of value to students and faculty as an IPM resource for advanced undergraduate- and graduate-level courses in IPM,

and for courses that examine alternative IPM systems. We also believe that the text will be useful to IPM practitioners, extension and outreach specialists and industry colleagues worldwide who have responsibilities for implementing sustainable IPM programs and policies. There is also much here that should be of interest to an audience of those concerned with a broad range of issues relating to agriculture production and/or environmental protection.


We thank Dr. Mark Ascerno, our Department Head at the University of Minnesota, for encouraging us to undertake this project. We thank Betty Radcliffe, surprisingly still married to the senior editor, for her many hours of expert copy-editing. We offer particular thanks to the 90 authors whose contributions and considerable expertise made

this work a reality. That these individuals not only contributed their time, talent, and insight, but also shared our enthusiasm for this project is greatly appreciated. Lastly, we thank Jacqueline Garget for proposing this project, and for her guidance and patience as we brought it to fruition.

Chapter 1

The IPM paradigm: concepts, strategies and tactics Michael E. Gray, Susan T. Ratcliffe and Marlin E. Rice Pests compete with humans for food, fiber and shelter and may be found within a broad assemblage of organisms that includes insects, plant pathogens and weeds. Some insect pests serve as vectors of diseases caused by bacteria, filarial nematodes, protozoans and viruses. Densities of many pests are regulated by density-independent factors, particularly under fluctuating environmental extremes (e.g. temperature, precipitation). Biotic components within a pest’s life system also may serve as important population regulation factors, such as interactions with predators and parasitoids. Some ecologists have theorized that competition (interspecific and/or intraspecific) for resources ultimately limits the densities and distributions of organisms, including those that are anthropocentrically categorized as pests.

1.1 Historical perspectives Humans have been in direct competition with a myriad of pests from our ancestral beginnings. Competition with pests for food intensified when humans began to cultivate plants and domesticate animals at the beginnings of agriculture, 10 000 to 16 000 years ago (Perkins, 2002; Thacker, 2002; Bird, 2003). As humans became more competent in producing crops used for food and fiber,

human densities began to increase and were organized in larger groupings such as villages. This increased concentration of humans in close proximity to their livestock is believed to have facilitated the mutation and spread of diseases across species in some instances. The earliest attempts at agricultural pest control were likely very direct and included handpicking and crushing insects, pulling or cutting weeds and discarding rotting food sources. Some pest control activities were inadvertent and included rotation or movement of crops (primarily planting crops in more fertile areas) and selection of plants for seed that had the greatest yields for sowing the following growing season. The reasoned use of pesticides is centuries old (2500 BC) dating back to when sulfur was directed at the control of mites and insects (Bird, 2003; Kogan & Prokopy, 2003). The ancient Egyptians also are credited with the use of compounds extracted from plants to aid in the control of insects and approximately 2000 years ago, Pliny listed arsenic and olive oil as pesticides. (Thacker, 2002). In AD 307, biological control was utilized in Chinese citrus orchards (Bird, 2003) and in AD 1100 soap was being used as an insecticide in China (Kogan & Prokopy, 2003). Perkins (2002) asserted that pest control began to transform significantly about four centuries ago:

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



About 400 years ago in Western Europe, a set of transformations completely changed economic life and, with it, pest control. New machines and new ways of making metals enabled industrialization. The new industrial processes were themselves linked to a new philosophy of nature, in which humans learned to manipulate natural processes more powerfully, particularly energy resources.

Thacker (2002) provides a list of insecticidal plants and their active compounds discovered by Europeans following the sixteenth century: sabadilla (Sabadilla officinarum) (c. 1500s); nicotine (Nicotiana tabacum) (late 1500s); quassin (Quassia amara) (late 1700s); heliopsin (Heliopsis longipes) (early 1800s); ryanodine (Ryania speciosa) (1940s); naphthoquinones (Calceolaria andina) (1990s); and derris (Derris chinensis) (mid-1990s). Many of these insecticidal plants were already being used for pest control purposes by native cultures prior to European exploration of the New World (Thacker, 2002). In the late 1800s, inorganic compounds were discovered that offered impressive insecticidal and fungicidal properties. In 1865, the Colorado potato beetle (Leptinotarsa decemlineata) was controlled by Paris green (cupric acetoarsenite), the first synthetic insecticide (Metcalf, 1994). Prior to the introduction of potatoes by settlers (1850s) into the western plains of the USA, this beetle fed primarily on the buffalo burr (Solanum rostratum). This insect soon found potatoes to be an excellent host. Lead arsenate replaced Paris green and was used extensively for Colorado potato beetle control until DDT became more readily available. Plant pathologists also determined (1880s) that synthetic compounds such as Bordeaux mixture (copper sulfate and hydrated lime) reduced the severity of downy mildew in grape vineyards (Perkins, 2002). In subsequent years, other metabolic inhibitory fungicidal compounds were utilized, such as those containing mercury. Weed control was largely dependent upon plowing and hoeing until the introduction (early 1940s) of 2,4dichlorophenoxy acetic acid (Perkins, 2002). In addition to these early chemical approaches to pest control, farmers relied upon their rudimentary knowledge (Webster, 1913) of pest life cycles and the use of cultural tactics to limit crop losses. In 1939, the insecticidal properties of DDT (dichlorodiphenyltrichloroethane) were discov-

ered by Paul Herman M¨ ueller, a scientist with the Geigy Chemical Company. Most entomologists view this development as the beginning of the modern insecticide era. The pest control benefits of this new insecticide were regarded initially as miraculous. Some referred to DDT as the “wonder” insecticide (Metcalf, 1994). During World War II, DDT was used extensively to prevent epidemics of several insect-vectored diseases such as yellow fever, typhus, elephantiasis and malaria. The use of DDT for insect control in the production of crops, protection of livestock, in forestry and in urban and public health arenas soared in the late 1940s and 1950s. In 1946, DDT-resistant strains of the house fly (Musca domestica) were reported in Sweden and Denmark (Metcalf, 1994). Despite this “chink” in the armor of DDT, the promise of chemicals to deliver economical and effective pest control (including that of plant diseases and weeds) heralded in an atmosphere characterized by an over-reliance on pesticides throughout the 1950s and 1960s. This over-reliance on insecticides soon led to many significant ecological backlashes such as insecticide resistance, concentration of chlorinated hydrocarbon insecticides in the food chain, significant declines in densities of natural enemy (predators and parasitoids) populations, secondary outbreaks of pests, resurgence of primary pests and unwanted insecticide residues on fruits and vegetables. Critics of this over-reliance on pesticides argued that basic biological research on pest ecology and alternative management strategies were being ignored. Entomologists engaged in biological control efforts in California, cotton production in North and South America and production of fruit in orchard systems (Canada, Europe and the USA) were among the first to recognize many of the acute ecological problems associated with indiscriminate pesticide use (Kogan, 1998).

1.2 Early conceptual efforts in IPM development In 1959, University of California entomologists at Berkeley, Vernon Stern, Ray Smith, Robert van den Bosch and Kenneth Hagen, published a seminal paper entitled “The Integration of Chemical and


Biological Control of the Spotted Alfalfa Aphid.” In this paper, they offered the following statement concerning the integrated control concept: Whatever the reasons for our increased pest problems, it is becoming more and more evident that an integrated approach, utilizing both biological and chemical control, must be developed in many of our pest problems if we are to rectify the mistakes of the past and avoid similar ones in the future.

Many terms and concepts, now well known by entomologists, plant pathologists, weed scientists and IPM practitioners, were defined by these authors such as economic threshold, economic injury level and general equilibrium position. The following definitions are provided from Stern et al. (1959): economic injury level economic threshold

general equilibrium position

The lowest population density that will cause economic damage. The density at which control measures should be determined to prevent an increasing pest population from reaching the economic injury level The average density of a population over a period of time (usually lengthy) in the absence of permanent environmental change

Integrated control was defined as applied pest control which combines and integrates biological and chemical control and employed the use of economic thresholds to determine when chemical control should be utilized to prevent pests from reaching the economic injury level. The integrated control concept has evolved into the IPM concept that includes insects, plant pathogens, weeds and vertebrate pests. Since the initial tenets of the integrated control concept were developed in response to insect pests, not all of the early basics fit well with regard to the practical management of weeds, plant pathogens and vertebrate pests. Knake & Downs (1979) indicated that IPM should be an interdisciplinary approach rather than simply combining various control options within one discipline: “Weeds harbor insects and diseases,

diseases may kill insects and weeds, and insects can be used to control other insects and weeds.” Ford (1979) described three threshold types for plant pathology IPM programs: (1) a threshold addressing detection, (2) a threshold for prevention due to zero injury tolerance and (3) the more standardized economic injury threshold. Integrated vertebrate pest control applies ecology and only supports destruction of individual vertebrates as a last option to address animal damage (Timm, 1979). The impact of pest management implementation requires careful examination of potential benefits, costs and risks. While increased producer productivity is often considered a benefit, if it is obtained at a high environmental cost, the true economic impact may be obscured (Carlson & Castle, 1972). Higley & Wintersteen (1992) suggested that the traditional use of economic thresholds and injury levels are insufficient in estimating the hidden environmental externalities associated with insecticide use. Some debate persisted among academics throughout the 1960s and into the 1980s regarding the perceived differences between “pest management” and “integrated control” (Kogan, 1998). Smith & Reynolds (1966) presented the concept of integrated pest control as a multifaceted, flexible, evolving system that blends and harmonizes control practices in an organized way. They believed the system must integrate all control procedures and production practices into an ecologically based system approach aimed at producing high quality products in a profitable manner. While this debate ensued, Rachel Carson published Silent Spring in 1962. This book galvanized sentiment among the general public against the abuses of pesticide applications. She was criticized by some for her use of emotionally charged passages such as (from Chapter 3, “Elixirs of Death”): For the first time in the history of the world, every human being is now subjected to contact with dangerous chemicals, from the moment of conception until death. In the less than two decades of their use, the synthetic pesticides have been so thoroughly distributed throughout the animate and inanimate world that they occur virtually everywhere.

She is given deserved credit for inspiring a generation of environmentalists and forcing the scientific community and governmental agencies




to more closely scrutinize pesticide use and registration requirements. Eight years after Silent Spring was published, the US Congress mandated that the administration and enforcement of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) be transferred from the US Department of Agriculture (USDA) to a newly created federal entity, the US Environmental Protection Agency (EPA). The passage of FIFRA amendments over the past 30 years has resulted in policies aimed at reducing environmental and human health and safety risks that are linked with pesticide use (Gray, 2002). Kogan (1998) indicated the following with respect to the popularization of IPM: Not until 1972, however, were “integrated pest management” and its acronym IPM incorporated into the English literature and accepted by the scientific community. A February 1972 message from President Nixon to the US Congress, transmitting a program for environmental protection, included a paragraph on IPM.

Kogan (1998) further added that broad agreement had by then been reached on several key points regarding IPM: (1) integration meant the harmonious use of multiple methods to control single pests as well as the impacts of multiple pests; (2) pests were any organism detrimental to humans, including invertebrate and vertebrate animals, pathogens, and weeds; (3) IPM was a multidisciplinary endeavor; and (4) management referred to a set of decision rules based on ecological principles and economic and social considerations.

Some continue to debate the definition of IPM; however, the key components of this concept can be found in these four elements. More recently, in response to a national review of the federally supported US IPM Program (US General Accounting Office, 2001), and considerable stakeholder input, the USDA developed the “IPM Road Map” (see Chapter 37) with the ultimate objective of increasing IPM implementation by practitioners such as “land managers, growers, structural pest managers, and public and wildlife health officials.” The IPM Road Map (2003) offers a definition of IPM that includes the historical elements of IPM reviewed by Kogan (1998), and in many ways

extends the concept to focus on reducing the risks of economic and environmental losses. Within the IPM Road Map (May, 2004 version) IPM is defined as: . . . a long-standing, science-based, decision-making process that identifies and reduces risks from pests and pest management related strategies. It coordinates the use of pest biology, environmental information, and available technology to prevent unacceptable levels of pest damage by the most economical means, while posing the least possible risk to people, property, resources, and the environment. IPM provides an effective strategy for managing pests in all arenas from developed residential and public areas to wild lands. IPM serves as an umbrella to provide an effective, all encompassing, low-risk approach to protect resources and people from pests.

1.3 Kinds of pests The selection of a strategy and components of an IPM program are largely influenced by the status of a pest in relationship to its host. Four pest types are commonly recognized by IPM practitioners: (1) subeconomic, (2) occasional, (3) perennial and (4) severe (Pedigo & Rice, 2006). (1) The general equilibrium position of a subeconomic pest is always below the economic injury level, even during its highest population peaks. An insect in this category may cause direct losses but if the host (crop) values are modest, and the pest densities are always low, then it is not appropriate to initiate control practices whose costs exceed the value of host damage. (2) The general equilibrium position of an occasional pest is nearly always below the economic injury level but occasionally population peaks exceed this level. The occasional pest is a very common type of pest. It may be present on or near a host nearly every year, but only sporadically does it cause economic damage. (3) The general equilibrium position of a perennial pest is below the economic injury level but peak populations occur with such frequency that economic damage usually occurs yearly.


(4) A severe pest has a general equilibrium position that is always above the economic injury level so that when they occur in or on a host, economic damage is always the end result. As might be expected, perennial and severe pests cause the most serious damage and difficult challenges in an IPM program.

1.4 Pest management strategies and tactics A pest management strategy is the total approach to eliminate or reduce a real or perceived pest problem. The development of a particular strategy will be greatly influenced by the biology and ecology of the pest and its interaction with a host or environment. The goal should be to reduce pest status when addressing problems using pest management. Because both the pest and host determine pest status, modification of either or both of these may be emphasized in a management program. Therefore, four types of strategies (Pedigo & Rice, 2006) could be developed based on pest characteristics and economics of management: (1) do nothing, (2) reduce pest numbers, (3) reduce host susceptibility to pest injury and (4) combine reduced pest populations with reduced host susceptibility. Once a pest management strategy has been developed, the methods of implementing the strategy can be chosen. These methods are called pest management tactics, and several tactics may be used to implement a management strategy.

1.4.1 Do-nothing strategy All pest injury does not cause an economic loss to a host. Many hosts, especially plants and occasionally animals, are able to tolerate small amounts of injury without suffering economic damage. It is not uncommon for trivial insect injury to be mistaken for economically significant injury. This is most likely to occur when the pest population density is not considered in relationship to an economic threshold. If the pest density is below the economic threshold, then the do-nothing strategy is the correct approach; otherwise money would be expended on control that would not result in a net benefit. The do-nothing strategy is frequently

used when insects cause indirect injury to a host, or when a successful pest management program reduces the pest population and only surveillance of the remaining population is necessary. No tactics are used in the do-nothing strategy, but this does not imply that no effort is necessary or that pest suppression is not occurring. Sampling of the pest population is required to determine that the do-nothing strategy is the appropriate response, and environmental influences may reduce the population, resulting in pest suppression.

1.4.2 Reduce pest numbers Reducing pest densities to alleviate or prevent problems is probably the most frequently used strategy in pest management. This strategy is often employed in a therapeutic manner when populations reach the economic threshold or in a preventive manner based on historical problems (Pedigo & Rice, 2006). Two objectives may be desirable in attempting to reduce pest densities. If the pest’s long-term average density, or general equilibrium position, is low compared with the economic threshold, then the best approach would be to diminish the population peaks of the pest. This action would not appreciably change the pest’s general equilibrium position, but it should prevent damage from occurring during pest outbreaks. If, however, the pest population’s general equilibrium position is near or above the economic threshold, then the general equilibrium position must be lowered so that the highest peak populations never reach the economic threshold. This may be done by either reducing the carrying capacity of the environment, or by reducing the inherited reproductive and/or survival potential of the population (Pedigo & Rice, 2006). There are many tactics that can be used to reduce pest numbers including resistant hosts, insecticides, pheromones, mechanical trapping, natural enemies, insect growth regulators, release of sterilized insects and modification of the environment.

1.4.3 Reduce host susceptibility to pest injury One of the most environmentally compatible and effective strategies is to reduce host susceptibility to pest injury. This strategy does not modify the pest population; instead the host or host’s




relationship and interaction with the pest is changed to make it less susceptible to a potentially damaging pest population. A common form of this strategy is where plant cultivars or animal breeds are developed with a type of resistance, known as tolerance, which provides greater impunity to a pest than a similar plant or animal without the tolerance. The tolerance expressed by a plant or animal does not reduce the attacking pest population, but the injury caused by the pests has less of a detrimental affect on the host (i.e. yield loss in plants or weight loss in animals) than it does on a similar host without the tolerance. The other component to this strategy, ecological modification of factors that influence the distribution or abundance of a pest, also can reduce host susceptibility. Examples of this strategy would be reducing livestock exposure to a pest insect by moving them from an outdoor environment to an indoor facility or adjusting a crop planting date to create an asynchrony between a pest and a susceptible plant stage.

1.4.4 Combine reduced pest populations with reduced host susceptibility A strategy that combines the objectives of the previous strategies is a logical step in the development of a pest management program. A multifaceted approach is more likely to produce greater consistency than a single strategy using a single tactic. Experience has shown that a single strategy is more likely to fail when either, slowly or quickly, a single tactic approach falters. With the multifaceted approach, if one tactic fails, then other tactics operate to help modulate losses. The use of multiple strategies and tactics is the basic principle in developing an IPM program.

1.5 Funding IPM research and implementation Since the early 1970s, the USDA, the EPA and the National Science Foundation (NSF) have been the primary governmental agencies in the USA that have provided competitive and formula-based funding for research and extension IPM programs. The majority of these IPM research and extension

programs are conducted by investigators located at land-grant universities (Morrill Land-Grant Acts, 1862, 1890). Two of the most visible and comprehensive IPM pilot efforts included the Huffaker (1972–1979, $US 13 million in funding, EPA, NSF, USDA) and Adkisson (1979–1984, $US 15 million in funding, EPA, USDA) projects (Allen & Rajotte, 1990). The Huffaker Project concentrated on the development of IPM tactics for insect pests in cotton, soybeans, alfalfa, citrus fruits, and pome and stone fruits. The Adkisson Project expanded its range of targeted pests to include diseases, insects and weeds in alfalfa, apples, cotton and soybeans. In 1978, a USDA report from the Extension Committee on Organization and Policy recommended that $US 58 million be spent on extension IPM programs. This goal was never achieved and federal funding for extension IPM programs began to falter reaching approximately $US 7.0 million in the early 1980s (Allen & Rajotte, 1990). By 2006, federal formula funds [Smith-Lever 3(d)] allocated across the USA for extension IPM programs had risen to a modest $US 9.86 million, or roughly $US 200 000 per state. Reasons are diverse for the weakening political support and funding for new and large-scale IPM initiatives in the USA (Gray, 1995). These reasons include the perception that implementation of IPM would lead to greater overall reductions in pesticide use than has occurred in some cropping systems, political support for “older” programs often wanes over time in lieu of new initiatives, continued difficulty in quantifying successes and impact of IPM implementation, struggles of IPM leadership to clearly articulate the goals of IPM implementation, and increasing popularity of organic production practices. Funding of IPM research and implementation programs in developing countries is increasingly important as food production and environmental concerns intensify in many densely populated areas around the globe. Some key organizations and programs that fund and promote these IPM efforts include: Food and Agriculture Organization of the United Nations (FAO), United Nations Environment Program (UNEP) and the United Nations Development Program (UNDP). In 1995, the Global IPM Facility was established and is housed in FAO Headquarters in Rome, Italy. Co-sponsors of the Facility include FAO, UNEP,


UNDP and the World Bank (Stemerding & Nacro, 2003). It was hoped that the Facility would ultimately result in more lending operations that would support IPM implementation. Thus far, the impact of the Global IPM Facility has been assessed as “mixed” (Schillhorn van Veen, 2003). Other key organizations that fund and promote IPM globally include the Integrated Pest Management Collaborative Research Support Program (IPM CRSP). This program was started in 1993 with the financial assistance of the US Agency for International Development (USAID). Current sites include: Albania, Bangladesh, Ecuador, Guatemala, Jamaica, Mali, Philippines and Uganda. Several USA institutions (Virginia Tech, Ohio State University, Purdue University) provide personnel who collaborate with scientists at the host institutions. Successful IPM programs that have been developed through this effort include: rice and vegetable cropping systems in the Philippines, maize and bean cropping systems in Africa, horticultural export crops in Latin America and sweet potato production in the Caribbean (Gebrekidan, 2003). Significant international contributions in host plant resistance to a variety of pests in crops have been achieved through support of the Consultative Group on International Agricultural Research (CGIAR) centers. These centers support the implementation of systemwide programs on IPM in several international “target zones” such as Africa, Asia and Latin America (James et al., 2003).

1.6 Measuring IPM implementation Assessing the level of IPM implementation has historically presented a challenge to policy makers, governmental agencies and scientists (Wearing, 1988). In an era of increasing pressure to ensure accountability, continued governmental support of IPM programs (research and extension) is contingent upon documenting increasing levels of IPM adoption and proving impact (economic, environmental and human health and safety benefits). Not all scientists, policy makers or practitioners of IPM agree that the primary goal of IPM is to reduce

pesticide use (Gray, 1995; Ratcliffe & Gray, 2004). The US Council on Environmental Quality (1972) described IPM as “an approach that employs a combination of techniques to control the wide variety of potential pests that may threaten crops.” It suggests numerous economic pests can be managed by “maximum reliance” on natural pest controls with the incorporation of key elements including cultural methods, pest-specific diseases, resistant crop varieties, sterile insects and attractants together with the use of biological control and reduced risk, species-specific chemical controls as part of an IPM program. Risk management and the fear of crop loss is often overemphasized, but coupled with the lack of implementation incentives many producers choose to only adopt limited aspects of IPM rather than a whole system approach (US Council on Environmental Quality, 1972). In September 1993 (US Congress, 1993) the Clinton Administration set a goal for 75 percent implementation of IPM practices, by 2000, on managed agricultural areas in the USA. A National Agricultural Statistics Service (2001) report indicated that by 2000, IPM adoption levels for many crops had met or exceeded this goal. However, in 2001, the United States General Accounting Office (GAO) published a document that was critical of the coordination and management of federal IPM efforts (across more than a dozen federal agencies). In addition, some criticism within the GAO report was directed at the lack of measurement and evaluation tools (environmental and economic) for assessing the level of IPM implementation. Since 2000, four regional IPM Centers within the USA have sought to improve the coordination of IPM implementation efforts utilizing a National Road Map for IPM (first articulated at the 4th National IPM Symposium, Indianapolis, IN, April 2003; see Chapter 37) as a blueprint (Ratcliffe & Gray, 2004). Bajwa & Kogan (2003) provide a very good assessment of IPM adoption in Africa, Americas (other than USA), Asia, Australia, Europe and the USA for many crops. The percentage of farmers who have adopted IPM practices is very high in many cases, such as: pear production in Belgium (98 percent), cotton production in Australia (90 percent), pome fruits in British Columbia (75 percent), and sugarcane production in




Colombia (100 percent). Despite these advances in IPM implementation, pesticide usage has increased in many developing countries throughout the 1990s and remains the exclusive tactic to control pests. Bajwa & Kogan (2003) remind us that “IPM is a tangible reality in some privileged regions of the world, but still remains a distant dream for many others.”

1.7 Examples of successful implementation of IPM 1.7.1 Ecological management of environment: push–pull polycropping in Africa Push–pull strategies use a combination of behavior-modifying stimuli to manipulate the distribution and abundance of pest or beneficial insects in pest management with the goal of pest reduction on the protected host or resource (Cook et al., 2007). Pests are repelled or deterred away from the resource (push) by using stimuli that mask host apparency or are deterrent or repellant. Pests are simultaneously attracted (pull), using highly apparent and attractive stimuli, such as trap crops, where they are concentrated, facilitating their elimination (Cook et al., 2007). The most successful push–pull strategy was developed for subsistence farmers in east Africa. Maize (Zea mays) and sorghum (Sorghum bicolor) two principal foods in east Africa, are attacked by lepidopteran stem borers, e.g. Busseola fuscus, Chilo partellus, Eldana saccharina and Sesamia calamistis, that cause 10–50 percent yield losses (Cook et al., 2007). Farmers combine the use of intercrops and trap crops, using plants that are appropriate for the farmers and exploit natural enemies. Stem borers are repelled from the maize and sorghum by non-hosts such as greenleaf desmodium (Desmodium intortum), silverleaf desmodium (Desmodium uncinatum) and molasses grass (Melinis minutiflora), which are interplanted with the maize or sorghum (the push). Around the field edges are planted trap crops, mostly Napier grass (Pennisetum purpureum) and Sudan grass (Sorghum vulgare sudanense), which attract and concentrate the pests (the pull). These grasses have a dual purpose as they are also used as

forage for livestock. Molasses grass, as an intercrop, reduces stem borer populations by producing stem borer repellent volatiles; it also increases parasitism by a parasitoid wasp. Desmodium also produces similar repellent volatiles; but also produces sesquiterpenes that suppress the parasitic African witchweed (Striga hermonthica), a major yield constraint of cropland in east Africa. The desmodium compounds stimulate germination of witchweed seeds and subsequent mortality of the seedlings. The push–pull strategy has contributed to increased grain yields and livestock production in east Africa, resulting in significant impact on food security (Cook et al., 2007).

1.7.2 Biological control: prickly pear cactus and cactus moths in Australia Prickly pears, or prickly pear cactus (Opuntia spp.), are native to the Americas but have become serious invasive weeds in suitable habitats around the world. Around 1840, cuttings of prickly pears were brought to Queensland, Australia for use as a hedge around fields and homesteads, as a botanical curiosity, and for production of cochineal – a dark reddish dye produced by scale insects that feed on the plant. Livestock and native birds quickly spread prickly pear seeds across overgrazed grasslands, where competition was reduced during droughts, whereas during heavy rainfall, broken pieces of prickly pears were carried into the interior on westward-flowing rivers (DeFelice, 2004). The climate and soil of eastern Australia was ideal for prickly pear and the weed quickly spread. Attempts were made by farmers and ranchers in the 1880s to control the weed, but were without success. In 1893, it was declared a noxious weed in Queensland. By 1913, prickly pear was estimated to cover 1.4 million ha with dense infestations and another 4.9 million ha with scattered infestations. By 1926, the prickly pear had infested 24 million ha in Queensland and New South Wales and was spreading at the rate of 1 million ha annually (DeFelice, 2004). Attempts at controlling the prickly pear using mechanical, chemical and cultural methods completely failed to stop the spread of the weed, mostly because control was poorly supported and many government policies only conspired to worsen the problem (DeFelice, 2004).


The infestation was so dense the 12 million ha were rendered useless, resulting in worthless grazing land and the abandonment of many farms and homesteads. In 1927, hope appeared in the form of an imported parasitic insect from South America – the cactus moth (Cactoblastis cactorum). This insect was evaluated and confirmed to only feed on prickly pear. Over 220 million eggs were reared and distributed and three years later 200 000 ha of prickly pear were destroyed. The insect rapidly spread and by the end of 1931, millions of hectares of prickly pears were a mass of rotting vegetation (DeFelice, 2004). Land that had been useless for decades was cleared and restored to rangelands and agricultural production. The prickly pear experience in Australia was one of the most frightening cases in history of ecological destruction by an invasive plant and also one of the most successful biological control campaigns ever mounted against a pest.

1.7.3 Sterile insect technique: screwworm eradication in North and Central America The classic achievement of success with the sterile insect technique was the eradication of the screwworm (Cochliomyia hominivorax) from the USA, Mexico and Central America. The screwworm is an obligate parasite of livestock and has occasionally attacked humans. The adult fly lays up to 450 eggs in open wounds where the larvae feed on tissues and enlarge the wound (Krafsur et al., 1987). Feeding by the larvae attracts other flies to oviposit in the wound, thereby aggravating the damage to the animal. Heavily parasitized livestock may be killed within 10 days. Historical livestock losses to this pest were astronomical. Prior to the sterile release program, losses were estimated at $US 70–100 million annually across the southern USA from Florida to California. A severe pest outbreak occurred in this region in 1935, with 1.2 million cases of infestation and 180 000 livestock deaths. The sterile insect technique involves the intentional release of large numbers of sterilized insects to compete with wild insects for mates (Krafsur et al., 1987). The sterile insect technique with screwworms involves the mass rearing of larvae on a specialized liquefied diet of bovine

blood and powdered egg. The pupae are collected from the rearing containers and at five days of age are irradiated with cesium. Female flies irradiated with this process fail to undergo vitellogenesis and therefore do not deposit eggs. Male flies likewise are sterilized and when they mate with a wild-type female, no viable eggs are produced. The concept of the sterile insect technique was put to the test in a pilot program on Sanibel Island, Florida and produced positive results. A larger test was initiated in 1954 on Curac¸ao, an 444-km2 island off the coast of Venezuela, where 400 sterilized males per 2.6 km2 were released for three months. The effort resulted in the complete eradication of the screwworm from the island and demonstrated the potential of the technique. The technique was then applied to livestock in Florida and southern Georgia and Alabama in 1958. More than 2000 million sterilized flies were released from airplanes during an 18-month period, resulting in complete eradication from the region. The program was then moved to southwestern USA in the early 1960s where sterile flies were released along the international border with Mexico. This resulted in a fly-free zone nearly 3200 km long and 500 to 800 km deep which prevented the flies from moving north into the USA. Fly infestation reports dropped from more than 50 000 in 1962 to 150 by 1970. Unfortunately, infestations did not remain low so a cooperative agreement between the USA and Mexican governments worked together to push the screwworm further south in Mexico. By 1986, Mexico was declared free of screwworm. The fly-free zone was continually moved south, eradicating the pest from numerous Central American countries. A fly-free barrier is currently maintained in Panama to prevent reinfestations from South America. In 1992, Raymond Bushland and Edward Knipling received the World Food Prize for their collaborative achievements in developing the sterile insect technique for eradicating or suppressing the threat posed by pests to crops and livestock.

1.7.4 Transgenic plants: control of European corn borer in North America The European corn borer (Ostrinia nubilalis) has been considered by some (Ostlie et al., 1997) to be the most damaging pest of maize in North




America with damage and control costs exceeding $US 1000 million during the early to mid-1990s. Insecticides were occasionally used by growers to prevent stalk tunneling, kernel damage and fallen ears in maize but often they were reluctant to embrace chemical control (Rice & Ostlie, 1997). Reasons for reluctance included the fact that larval damage was hidden, large infestations are unpredictable, fields had to be scouted multiple times requiring time and skill, insecticides were expensive and raised environmental and health concerns and benefits of insecticide control were uncertain. These concerns paved the way for a novel way of managing this pest through the use of transgenic plants. In 1996, Mycogen Seeds and Novartis Seeds introduced the first commercial Bt maize hybrids. The Bt hybrids were genetically transformed to express a gene from the soil bacterium, Bacillus thuringiensis, which produces a protein that is toxic to European corn borer larvae. Most larvae die after taking only a few bites of maize leaf tissue. Consequently, Bt maize provides extremely high levels of larval mortality resulting in exceptional yield protection even during heavy infestations of European corn borer (Ostlie et al., 1997). In 2005, approximately 35 percent of the maize hectares were planted to a corn borer resistant transgenic hybrid with the result being that during the past ten years, the European corn borer has had a steady decline in the severity of populations, thereby leading some to conclude that the insect has become a secondary pest (Gray, 2006). An additional effect was that the percent of farmers who decreased their insecticide use doubled during the first three years of planting a transgenic maize hybrid resulting in less broad-spectrum insecticide applied to the fields (Pilcher et al., 2002). Maize growers perceive that less exposure to insecticides and less insecticide in the environment are the two primary benefits of planting transgenic maize hybrids (Wilson et al., 2005). The success of commercial transgenic Bt maize has lead to the development of triplestacked hybrids that may express a protein for corn borers, a different protein specific for corn rootworms (Diabrotica spp.) and resistance to herbicides.

1.7.5 Insect growth regulators: termite control in North America Termites are destructive pests of wooden structures and the latest industry estimates place the annual cost of damage and treatment at $US 5000 million worldwide (National Pest Management Association, 2005). Termite control generally consists of five types of treatment programs: liquid termiticides, bait systems, wood preservatives, mechanical barriers and biological termiticides (Hu, 2005). Each type of program has its advantages and disadvantages, but the bait system is the most novel as it uses an insect growth regulator to control the termite colony. The bait system is a relatively new tool for termite control. Instead of applying a chemical barrier designed to exclude termites from a wooden structure, termites are offered food in the form of baits (Hu et al., 2001). Treatment baits have two components: a termite food source, such as a block of wood in the soil, and a slow-acting termiticide, often an insect growth regulator. The insect growth regulator (diflubenzuron, hexaflumuron or noviflumuron) is a slow-acting, nonrepellent toxicant that prevents the formation of chitin in the insect cuticle. Termites feeding on the bait are not killed immediately, but through colony recruitment when worker termites find the bait the insect growth regulator is passed to other colony members, ultimately leading to decline or perhaps elimination of the colony. The advantage of baiting is that the system is non-intrusive, consumer friendly, safer than most of the soil-applied insecticides, specifically targets termites and dramatically reduces the amount of chemical needed to protect a structure. However, a disadvantage is that the process may take weeks or months to knock down termite populations.

1.8 IPM within a transgenic era In 1996, transgenic crops were commercialized on a limited basis in the USA for the first time. In ten years, the use of transgenic crops has seemingly transformed the IPM paradigm, particularly in the major field crops arenas. The primary transgenic tools include the planting of


herbicide-tolerant varieties of cotton, maize and soybeans (primarily to the herbicides glyphosate and glufosinate-ammonium) as well as Bt cotton and maize that express proteins derived from various strains of the bacterium Bacillus thuringiensis. Fernandez-Cornejo et al. (2006) reported that 87 percent and 60 percent of soybean and cotton hectares, respectively, in the USA were planted to herbicide-tolerant varieties in 2005. The use of Bt cotton and maize also was impressive, estimated at 52 percent and 35 percent, respectively, on USA hectares in 2005. Although the USA is estimated to account for 55 percent of the global area devoted to transgenic crops, 21 other nations in 2005 were planting transgenic crops as well. Nations that are heavily engaged in the production of transgenic crops (James, 2005) include the USA (49.8 million ha), Argentina (17.1 million ha), Brazil (9.4 million ha), Canada (5.8 million ha) and China (3.3 million ha). Increasingly the dialogue among scientists engaged in IPM research and extension programs in field crops has turned towards resistance management. For many of these scientists, their focus has shifted to that of evaluating models and recommending to producers the best deployment of transgenic crops across the agricultural landscape to delay the onset of resistance to these new tools (Gould, 1998). The philosophical debate will continue to rage for many years regarding the “fit” of transgenic crops within the IPM framework. For supporters, transgenic crops fit within the host plant resistance pillar of IPM. For others, transgenic crops, such as Bt maize and cotton, are little different than using a broadcast insecticide application on a prophylactic basis. While global adoption rates of transgenic crops are expected to increase, the philosophical debate within the academic community will continue on what truly constitutes IPM.

References Allen, W. A. & Rajotte, E. G. (1990). The changing role of extension entomology in the IPM era. Annual Review of Entomology, 35, 379–397. Bajwa, W. I. & Kogan, M. (2003). Integrated pest management adoption by the global community. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia,

D. Dakouo & D. Mota-Sanchez, pp. 97–107. Wallingford, UK: CABI Publishing. Bird, G. W. (2003). Role of integrated pest management in sustainable development. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia, D. Dakouo & D. Mota-Sanchez, pp. 73–85. Wallingford, UK: CABI Publishing. Carlson, G. A. & Castle, E. N. (1972). Economics of pest control. II. Systems approach to pest control. In Pest Control Strategies for the Future, pp. 79–99. Washington, DC: National Academy of Sciences. Carson, R. L. (1962). Silent Spring. Boston, MA: Houghton Mifflin Company. Cook, S. M., Khan, Z. R. & Pickett, J. A. (2007). The use of push–pull strategies in integrated pest management. Annual Review of Entomology, 52, 375–400. DeFelice, M. S. (2004). Prickly pear cactus, Opuntia spp.: a spine-tingling tale. Weed Technology, 18, 869– 877. Fernandez-Cornejo, J., Caswell, M., Mitchell, L., Golan, E. & Kuchler, F. (2006). The First Decade of Genetically Engineered Crops in the United States, USDA Economic Information Bulletin No. 11. Washington, DC: US Government Printing Office. Ford. R. E. (1979). The role of plant pathology in integrated pest management. In Integrated Pest Management North Central Region Workshop Proceedings, St. Louis, MO, December 11–13, 1979, ed. S. Elwynn Taylor, Section I, pp. 23–33. Gebrekidan, B. (2003). Integrated pest management collaborative research support program (USAID–IPM CRSP): highlights of its global experience. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia, D. Dakouo & D. Mota-Sanchez, pp. 407–418. Wallingford, UK: CABI Publishing. Gould, F. (1998). Sustainability of transgenic insecticidal cultivars: integrating pest genetics and ecology. Annual Review of Entomology, 43, 701–726. Gray, M. E. (1995). Status of CES–IPM programs: results of a national IPM coordinators survey. American Entomologist, 41, 136–138. Gray, M. E. (2002). Federal Insecticide, Fungicide, and Rodenticide Act. In Encyclopedia of Pest Management, ed. D. Pimentel, pp. 261–262, New York: Marcel Dekker. Gray, M. (2006). European corn borer: a secondary pest for now? Pest Management and Crop Development Bulletin, 9, 261–262. Available at bulletin/article.php?id=530. Higley, L. G. & Wintersteen, W. K. (1992). A novel approach to environmental risk assessment of pesticides as a basis for incorporating environmental




costs into economic injury levels. American Entomologist, 38(2), 34–39. Hu, X. P. (2005). Subterranean Termite Control Products for Alabamians. Alabama Cooperative Extension System, Auburn University. Available at docs/A/ANR-1252/. Hu, X. P., Appel, A. G., Oi, F. M. & Shelton, T. G. (2001). IPM Tactics for Subterranean Termite Control, ANR-1022. Alabama Cooperative Extension System, Auburn University. Available at ANR-1022/. James, B., Neuenschwander, P., Markham, R. et al. (2003). Bridging the gap with the CGIAR systemwide program on integrated pest management. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia, D. Dakouo & D. Mota-Sanchez, pp. 419–434. Wallingford, UK: CABI Publishing. James, C. (2005). Executive summary. In Global Status of Commercialized Biotech/GM Crops: 2005, ISAAA Briefs No. 34. Ithaca, NY: International Service for the Acquisition of Agri-biotech Applications. Knake, E. L. & Downs, J. P. (1979). The weed science phase of pest management. In Integrated Pest Management North Central Region Workshop Proceedings, St. Louis, MO, December 11–13, 1979, ed. S. Elwynn Taylor, Section I, pp. 33–37. Kogan, M. (1998). Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology, 43, 243–270. Kogan, M. & Prokopy, R. (2003). Agricultural entomology. In Encylopedia of Insects, eds. V. H. Resh & R. T. Card´e, pp. 4–9. San Diego, CA: Academic Press. Krafsur, E. S., Whitten, C. J. & Novy, J. E. (1987). Screwworm eradication in North and Central America. Parasitology Today, 3, 131–137. Metcalf, R. L. (1994). Insecticides in pest management. In Introduction to Insect Pest Management, 3rd edn, eds. R. L. Metcalf & W. H. Luckmann, pp. 245–314. New York: John Wiley. National Agricultural Statistics Service (2001). Pest Management Practices: 2000 Summary, Sp Cr 1 (01). Washington, DC: US Department of Agriculture. National Pest Management Association (2005). The Big Bite of Termites: $5 Billion a Year in Damages. Available at National Roadmap for Integrated Pest Management (2003). In Proceedings, Integrated Pest Management for Our Environment – for Our Future, 4th National Integrated Pest Management Symposium, Indianapolis, IN, April 8–10, 2003, pp. 9–11. Urbana, IL: University of Illinois. Ostlie, K. R., Hutchison, W. D., Hellmich, R. L. et al. (1997). Bt-Corn and European Corn Borer: Long-Term Suc-

cess Through Resistance Management. North Central Regional Extension Publication NCR 602. St. Paul, MN: University of Minnesota. Pedigo, L. P. & Rice, M. E. (2006). Entomology and Pest Management, 5th edn. Upper Saddle River, NJ: Pearson Prentice Hall. Perkins, J. H. (2002). History. In Encyclopedia of Pest Management, ed. D. Pimentel, pp. 368–372. New York: Marcel Dekker. Pilcher, C. D., Rice, M. E., Higgins, R. A. et al. (2002). Biotechnology and the European corn borer: measuring historical farmer perceptions and the adoption of transgenic Bt corn as a pest management strategy. Journal of Economic Entomology, 95, 878–892. Ratcliffe, S. T. & Gray, M. E. (2004). Will the USDA IPM centers and the national IPM roadmap increase IPM accountability? – responses to the 2001 General Accounting Office report. American Entomologist, 50(1), 6–9. Rice, M. E. & Ostlie, K. R. (1997). European corn borer management in field corn: a survey of perceptions and practices in Iowa and Minnesota. Journal of Production Agriculture, 10, 628–634. Schillhorn van Veen, T. W. (2003). The World Bank and pest management. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia, D. Dakouo & D. Mota-Sanchez, pp. 435–440. Wallingford, UK: CABI Publishing. Smith, R. F. & Reynolds, H. T. (1966). Principles, definitions and scope of integrated pest control. In Proceedings of the FAO Symposium on Integrated Pest Control, Rome, Italy, October 11–15, 1965, 1, 11–18. Rome, Italy: Food and Agriculture Organization of the United Nations. Stemerding, P. & Nacro, S. (2003). FAO integrated pest management programs: experiences of participatory IPM in West Africa. In Integrated Pest Management in the Global Arena, eds. K. M. Maredia, D. Dakouo & D. Mota-Sanchez, pp. 397–406. Wallingford, UK: CABI Publishing. Stern, V. M., Smith, R. F., van den Bosh, R. & Hagen, K. S. (1959). the integration of chemical and biological control of the spotted alfalfa aphid (the integrated control concept). Hilgardia, 29(2), 81–101. Thacker, J. R. M. (2002). An Introduction to Arthropod Pest Control. Cambridge, UK: Cambridge University Press. Timm, R. M. (1979). Vertebrate zoology’s role in integrated pest management. In Integrated Pest Management North Central Region Workshop Proceedings, St. Louis, MO, December 11–13, 1979, ed. S. Elwynn Taylor, Section I, pp. 7–15.


US Congress (1993). Testimony of Carol M. Browner, Administrator EPA; Richard Rominger, Deputy Secretary of Agriculture; and David Kessler, Commissioner of FDA. Hearings before the Committee on Labor and Human Resources, US Senate, and Subcommittee on Health and the Environment, Committee on Energy and Commerce, US House of Representatives, 22 September, 1993. US Council on Environmental Quality (1972). Integrated Pest Management. No. 4111–0010, pp. 9–15. Washington, DC: US Government Printing Office. US General Accounting Office (2001). Agricultural pesticides: management improvements needed to further promote

integrated pest management. Report GAO-01–815. Washington, DC: US Government Printing Office. Wearing, C. H. (1988). Evaluating the IPM implementation process. Annual Review of Entomology, 33, 17–38. Webster, F. M. (1913). Bringing applied entomology to the farmer. In Department of Agriculture Yearbook, pp. 75–92. Washington, DC: US Government Printing Office. Wilson, T. A., Rice, M. E., Tollefson, J. J. & Pilcher, C. D. (2005). Transgenic corn for control of the European corn borer and corn rootworms: a survey of Midwestern farmer practices and perceptions. Journal of Economic Entomology, 98, 237–247.


Chapter 2

Economic impacts of IPM Scott M. Swinton and George W. Norton Economic impact analyses of IPM programs measure the economic effects on producers and consumers that can be attributed to IPM programs and practices. Good impact assessments are tailored to the objectives of the programs they are evaluating. Because IPM program objectives and approaches can vary widely, there is no one-sizefits-all method for IPM impact assessment. Some IPM activities are narrowly focused, such as a new methodology for measuring pest density. Others are broad, such as a national training program in pest recognition. Small programs may have narrow impacts, while large programs may have repercussions great enough to change prices at the regional or even national level. Despite the diversity of approaches and objectives, virtually all IPM programs aim to influence economic and health or environmental outcomes. Economic outcomes may be measured at the level of the individual management unit (e.g. a farm) or at the aggregate level of all producers and consumers in a given market. Environmental and health outcomes may be measured using indicative, average approaches or using site-specific data about environmental vulnerability. Decision makers often wish to explore the trade-offs between economic and environmental/health outcomes and methods exist for that purpose. This chapter offers an overview of economic impact analysis methods, including

ways to incorporate environmental and health effects; in part this chapter also draws upon previous summaries of impact assessment methods by Norton & Mullen, 1994; Norton et al., 2001; Cuyno et al., 2005; and Norton et al., 2005. Recognizing that economic impact analysis is often performed under tight budget constraints, the methods proposed range from basic and somewhat unreliable but inexpensive, at one extreme, to nuanced and reliable, but costly, at the other extreme.

2.1 Measuring IPM adoption Impact assessments must measure and attribute impacts before the costs and benefits of those impacts can be calculated. The measurement of IPM impacts typically begins with estimating the extent of IPM adoption. Unlike simple technologies whose adoption might be measured by a yes/no question (e.g. planting seed of an improved crop variety), IPM often involves a suite of different practices that are not necessarily adopted jointly. Recognizing different degrees of IPM practice adoption is central to properly measuring IPM adoption. The way that the degree of adoption is incorporated will depend upon whether adoption is estimated by expert opinion or by survey methods.

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


Short of relying wholly on secondary data, the simplest, least costly approach to measuring IPM adoption is to rely upon expert opinion. For this purpose, IPM practices can be bundled together into ideal typical groupings to represent a range from no adoption to advanced IPM adoption. An expert-opinion-based estimate of IPM adoption might categorize adoption levels into: (1) no IPM, (2) weekly pest scouting, (3) weekly pest scouting with use of a threshold-based decision rule for treatment and (4) weekly pest scouting with a threshold-based decision rule and pesticide rates reduced by some specified level, say 50 percent. The chief limitation to relying upon expert opinion to estimate IPM adoption levels is that even the best experts often have a biased perspective. They are familiar with IPM adoption levels among those with whom they have contact. But they do not have contact with a random sample of the population; instead, they tend to meet individuals interested in IPM. One comparison of IPM adoption estimates by extension agent experts versus scientific survey results found that while the experts estimated that 98% of Michigan tart cherry producers had adopted some form of IPM by 1999, a randomly sampled farm survey found that only 64% had done so.1 For survey purposes, IPM adoption must be described in terms that are simple and explicitly additive. Whereas the ideal types of adopters used for expert opinion purposes may involve a detailed mix of practices at each level, real producers will constitute a continuum of adoption levels that is likely to include surprising combinations or absences of practices. Hence, surveys must base adoption measures on binary choices that are exhaustive. To contrast with the ideal type adoption categories above, a survey would need to rely on simpler categories such as: (1) no IPM, (2) scouted for pests at least one time, (3) scouted for pests and used a threshold decision rule at least once and (4) did (3) and sprayed with a median rate less than or equal to 50% of the label pesticide rate. Note each of these adoption levels is based on the presence or absence of a general condition, with


those conditions being additive (scouting, threshold rule, reduced rate), which permits them to be summed up to 100% of all producers. In addition to estimating the level of IPM adoption, impact assessments must sometimes be able to attribute adoption to a particular program. For an extension program, it will be important to show how much adoption can be attributed to that program, versus the amount that would have occurred anyway in its absence. Evaluators refer to this benchmark case as the “counterfactual,” that is, what would have occurred without the program. When impact assessments are planned from the outset, the attribution task can be facilitated in two ways. First, a benchmark survey before the program begins can establish the degree of adoption, as well as other baseline conditions. Second, a “no treatment” control area may be planned as a basis for comparison with the program intervention area. Such an arrangement explicitly builds in a counterfactual case that is subject to other evolutionary effects outside of the IPM program. When a no-treatment control area is not planned from the outset, the next best means to decide whether adoption is attributable to the program is to use a statistical regression model. The first step is to survey adopters and non-adopters. Using their responses, a statistical regression model is developed where the dependent variable is the decision whether or not to adopt (or some variant, such as the number of IPM practices adopted). The explanatory variables include all those that might explain adoption, including variables unrelated to the IPM program of interest. Economists typically include input and output prices, important producer and business traits, and local environmental characteristics, as well as a measure of IPM program exposure. Such a model can help to predict the degree of influence by the IPM program, when taking into account other factors that might also have influenced IPM adoption (McNamara et al., 1991; Fernandez-Cornejo, 1998). Those who participate in IPM programs may be different in some respects than those who do not

Unpublished data from a Michigan Agricultural Statistics Service survey conducted by S. M. Swinton in 1999.




Additions to Net Revenue

Reductions in Net Revenue

Increased Returns:

Decreased Returns:

1. __________ $______

4. __________ $______

2. __________ $______

5. __________ $______

3. __________ $______

6. __________ $______

Total $______ (A)

Fig. 2.1 Partial budget form.

Total $______ (B)

Decreased Costs:

Increased Costs:

7. __________ $______

10. __________


8. __________ $______

11. __________


9. __________ $______

12. __________


Total $______ (C)

Total $______ (D)

A+C = $______ (E)

B+D = $______ (F)

Change in Net Returns = E - F = $________

participate, and as a result the estimated effects of IPM exposure in the regression described above may be biased. To reduce the possibility of bias, a second regression can be run in which IPM program participation or exposure is regressed on a set of producer characteristics or other variables. The results of this regression can be used to generate a predicted IPM program exposure variable which is used in the adoption regression in place of the simple IPM exposure variable to remove the bias (Feder et al., 2004).

2.2 Individual economic impacts to the IPM user 2.2.1 Analyzing typical adoption cases using budgets Economic impacts are realized by individual IPM users (e.g. producers, households, schools) when they adopt IPM practices and programs. The level and variability of costs and/or returns may change with IPM. Enterprise budgets or partial budgets are commonly used to assess relative costs and returns of non-IPM versus IPM practices. An enterprise budget lists all estimated expenses and receipts associated with a particular enterprise,

while a partial budget only includes changes in yield, prices and costs to assess whether benefits (due to increased revenues and reduced costs) exceed burdens (from reduced revenues or increased costs). Budgets for various pest management alternatives can be compared using data from replicated experiments or from producer surveys. When using experimental data, enterprise budgets are commonly simplified to gross margin analysis, which includes only those costs that vary across treatments (Centro Internacional de Mejoramiento de Ma´ız y Trigo, 1988). Budget analysis assumes typical conditions, no carry-over effects from one period to the next, predictable prices, costs and yield effects of pests, and that the decision maker’s only objective is to maximize profits (Swinton & Day, 2003). Partial budgets are the most common and practical type of budget used for assessing IPM impacts. Budgets can be constructed for each adoption level (group of practices). A typical partial budget form is presented in Fig. 2.1. By developing a budget for each level of adoption, changes in net revenue can be associated with levels of IPM adoption. Data are required on inputs, outputs, and their prices. These data can be obtained from on-farm trials, user surveys or from taking existing budgets and adjusting the


cost categories. When based on replicated experiments, budget results such as gross margins can be subjected to analysis of variance to test for differences in mean profitability by treatment (Swinton et al., 2002). Results of budgeting analysis can be used by scientists and extension workers to judge the profitability of practices they are developing or recommending to farmers or of practices already adopted. A second major use of budget information is as an input into a market or societal level assessment of the economic benefits and costs of an IPM program as discussed below.

2.2.2 Analyzing optimal IPM decisions Economic evaluations of IPM programs at the user level may be concerned not only with the effects of IPM adoption, but also with the optimal level of pest control or the optimal IPM mix at a wholefarm scale. When profit maximization is the goal, optimal use of an IPM practice occurs when the marginal increase in net returns from applying another unit of the practice equals the marginal cost of its application. Entomologists and some weed scientists have applied this concept when identifying economic thresholds for pest densities (Pedigo et al., 1986; Cousens, 1987). An economic threshold is the pest population that produces incremental damage equal to the cost of preventing that damage. If pest density is below the threshold, no treatment is warranted. If it is above, treatment should occur to knock back pest density to below the economic threshold. IPM programs often involve scouting to inform producers about pest densities in relation to the threshold. Economic thresholds may be influenced by many factors including pesticide costs, output prices, the development of pesticide resistance, and the relationship between pest and predator levels and crop losses. If risk or environmental costs are considered, thresholds will be influenced by these factors as well (e.g. Szmedra et al., 1990). Optimal use of pest management practices can also be examined using mathematical programming techniques. For example, linear programming can be used to maximize an objective, such as net returns from a set of cropping activities that include IPM, subject to constraints on factors such as land, labor, capital, water and pesticide

runoff or residuals. Martin et al. (1991) provides an example, with an analysis of alternative tillage systems, crop rotations and herbicide use on eastcentral USA Corn Belt farms. An application of non-linear programming to a pest management problem related to pesticide resistance is found in Gutierrez et al. (1979). Dynamic programming allows for examination of optimal pest control strategies when time is included in the models, and variables such as plant product, pest population density and the stock of pest susceptibility to pesticides are functions of time. Zacharias & Grube (1986) provide an example of applying such a model to examine optimal control of corn rootworms (Diabrotica barberi and D. longicornis longicornis) and soybean cyst nematode (Heterodera glycines) in Illinois. Producers who consider adopting IPM strategies are often interested in the degree of risk as well as profitability. Risk may arise from biological, technical, or economic factors. The attractiveness of alternative pest management practices in the presence of risk can be assessed with a method called stochastic dominance (SD). With SD one can compare pairs of alternative pest management strategies for various sets of producers. These sets of producers are defined by their degrees of risk aversion. Examples of using SD in economic evaluation and comparison of IPM strategies with other strategies are found in Musser et al. (1981), Moffitt et al. (1983), and Greene et al. (1985); see also Chapter 4.

2.3 Aggregate or market-level economic impacts of IPM When many growers all adopt changed pest management methods, there may be non-linear impacts, such as changes in crop supplies that are large enough to affect prices. Likewise, the timing of impacts may affect their total value, as effects nearer in time are generally valued more than ones in the more distant future. Properly measuring the aggregate economic impacts of IPM across space and time requires some care to account for the possibility of price effects and discounted future values (Norton et al., 2005).




Price S0

P0 R


P1 d



c D





Fig. 2.2 IPM benefits measured as changes in economic surplus when supply shifts.

2.3.1 Capturing price effects when IPM impacts are large When widespread adoption of IPM occurs, changes in crop prices, cropping patterns, producer profits and societal welfare can occur. These changes arise because costs differ and because supplies may increase, affecting prices for producers and consumers. These changes are illustrated in Fig. 2.2. In this figure, S0 represents the supply curve before adoption of an IPM strategy, and D represents the demand curve. The initial price and quantity are P0 and Q0 . Suppose IPM leads to savings of R in the average and marginal cost of production, reflected in a shift down in the supply curve to S1 . This shift leads to an increase in production and consumption to Q1 (by Q = Q1 − Q0 ) and the market price falls to P1 (by P = P0 – P1 ). Consumers are better off because they can consume more of the commodity at a lower price. Consumers benefit from the lower price by an amount equal to their cost saving on the original quantity (Q0 × P) plus their net benefits from the gain in quantity consumed. Their total benefit is represented by the area P0 abP1. Although they may receive a lower price per unit, producers may be better off too, because 2

their cost has fallen by R per unit, an amount greater than the fall in price. Producers gain the increase in profits on the original quantity (Q0 × (R − P)) plus the profits earned on the additional output, for a total producer gain of P1 bcd. Total benefits are the sum of producer and consumer benefits. The distribution of benefits between producers and consumers depends on the size of the fall in price (P) relative to the fall in costs (R), the price elasticities of supply and demand, and on the nature of the supply shift. For example, if a commodity is internationally traded, production has less of an effect on price, so more benefits may to accrue to producers. Or, if the supply curve shifts in more of a pivotal fashion as opposed to a parallel fashion as shown in Fig. 2.3, the benefits to producers would be reduced. Examples of IPM evaluation using this type of model, often called the “economic surplus” model, are found in Napit et al. (1988). Formulas for calculating consumer and producer gains for a variety of market situations are found in Alston et al. (1995).2

2.3.2 Aggregating economic impacts over time Research and outreach programs in IPM often last many years, and their impacts may endure even longer. Benefit–cost analysis provides a framework for aggregating or projecting economic surplus over time by calculating net present values, internal rates of return or benefit–cost ratios. The benefits are the change in total economic surplus calculated for each year, and the costs are the expenditures on the IPM program. The primary purpose of benefit–cost analysis is to account for the fact that the sooner benefits and costs occur, the more they are worth. The net present value (NPV) of discounted benefits and costs can be calculated as follows: NPV =

T  Rt − C t (1 + i)t t=1


For example the formula to measure the total economic benefits to producers and consumers in Fig. 2.2, which assumes no trade, is KP0 Q0 (1 + 0.5Zn), where: K = the proportionate cost change, P0 = initial price, Q0 = initial quantity, Z = Ke/(e + n), e = the supply elasticity, and n = the demand elasticity. Other formulas would be appropriate for other market situations.


where: Rt = the return in year t T = (change in economic surplus) Ct = the cost in year t (the IPM program costs) i = the discount rate. A tool closely related to NPV is the internal rate of return (IRR), which measures the rate of return that would render benefits just equal to costs (the implied discount rate that would make NPV = 0).

2.4 Valuing health and environmental impacts Traditionally, reporting on changes in use of pesticide active ingredients associated with changes in pest management practices or just listing the number of pesticide applications have been common methods for assessing health and environmental (HE) impacts of IPM. However, specific pesticides and the means, timing and location of their application differentially affect health and the environment. Several methods have been developed to provide more refined evaluation of the effects of IPM on reducing HE risks. Some methods estimate monetary values associated with hazard reductions (Cuyno et al., 2005), while others present HE measures as trade-offs with profitability measures. Measuring HE benefits of IPM is difficult for various reasons. First, it is a challenge to assess the physical and biological effects of pesticide use that occur under different levels of IPM. Second, pesticides can have many distinct acute and longterm effects on sub-components of the health and the environment such as mammals, birds, aquatic life and beneficial organisms. Third, because the economic value associated with HE effects is generally not priced in the market, it is difficult to know how heavily to weight the various HE effects compared to one another and compared to profitability measures.

2.4.1 Hazard indexes and scoring models Location-specific models have been used that require detailed field information such as soil

type, irrigation system, slope and weather and produce information on the fate of chemicals applied. For example, the Chemical Environmental Index (CINDEX), based on another model GLEAMS, was developed by Teague et al. (1995) to describe the effects of pesticides on ground and surface water. CINDEX values are defined for individual pesticide use strategies. Calculations are based on the 96hour fish LC50 , lifetime Health Advisory Level (HAL) value, the US Environmental Protection Agency (EPA) Carcinogenic Risk Category and the runoff and percolation potential for each pesticide used in the strategy under consideration. Location specific models can be used to provide information on trade-offs between HE effects and income effects for pest management practices. Non-location-specific models have been more commonly used for impact assessment because they require less effort and data. These models require information on the pesticides applied and the method of application and produce indicators of risks by HE category as well as weighted total risk for the pesticide applications. Examples are the Environmental Impact Quotient (EIQ) developed by Kovach et al. (1992), the Pesticide Index (PI) of Penrose et al. (1994) and a multi-attribute toxicity index developed by Benbrook et al. (2002). Each indexing method is a type of scoring model that involves subjective weighting of risks across environmental categories. These methods perform two tasks (Norton et al., 2001): the first is to identify the risks of pesticides to the individual categories of health and the environment, such as groundwater, birds, beneficial insects and humans, and the second is to aggregate and weight those impacts across categories. The first task is complicated by the desire to identify mutually exclusive categories, especially ones with available data. The categories in most models contain a mixture of non-target organisms (e.g. humans, birds, aquatic organisms, beneficial insects, wildlife) and modes of exposure (e.g. groundwater, surface water). The second task is challenging because of the inherent subjectivity of the weights. A widely used non-location-specific index of HE impacts of pesticide use is the EIQ developed by Kovach et al. (1992). The EIQ uses a discrete ranking scale in each of ten categories to identify




a single rating for each pesticide active ingredient (a.i.). The categories include acute toxicity3 to non-target species (birds, fish and bees), acute dermal toxicity, long-term health effects, residue half-life (soil and plant surface), toxicity to beneficial organisms and groundwater/runoff potential. The EIQ groups the ten categories into three broad areas of pesticide action: farm worker risk, consumer exposure potential and ecological risk. The EIQ is then calculated as the average impact of a pesticide (AI) over these three broad areas and is reported as a single number. The EIQ is defined for specific pesticide active ingredients. In order to assess the actual damage from pesticide use on a specific field, the EIQ can be converted into a “field use rating.” If only one pesticide is applied, this rating is obtained by multiplying the pesticide’s EIQ by its percent a.i. and by the rate at which the pesticide was applied. Benbrook et al. (2002) developed an indexing method to monitor progress in reducing the use of high-risk pesticides. For pesticides used in Wisconsin potato production, multi-attribute toxicity factors were calculated that reflect each pesticide’s acute and chronic toxicity to mammals, birds, fish and small aquatic organisms and compatibility with bio-intensive IPM. These factors were multiplied by the pounds of active ingredients of the pesticides applied to estimate pesticide-specific toxicity units. These units can be tracked over time or related to use of IPM.

2.4.2 Monetary valuation of health and environmental effects One mechanism that can be used to reduce the subjectivity on the weights used in the EIQ or in studies that apply multi-attribute toxicity factors is to elicit information on individuals’ willingness to pay for risk reduction for the various HE components. There is usually no market price for reduced HE risk that can be used to provide these willingness-to-pay weights, but alternative monetary valuation techniques can be used such as contingent valuation (CV), experimental auctions and hedonic pricing (Champ et al., 2003; Freeman, 2003). These methods can be costly and 3

See Chapter 17 for details on toxicity indicators.

time consuming to perform well, so they are best executed by experts in non-market valuation. Perhaps the most direct market-based way to estimate the human health impact of environmental damage or illness is to measure the associated costs. Studies have been completed on the cost of medical treatment and productivity losses due to pesticides in the Philippines and Ecuador (Antle & Pingali, 1994; Pingali et al., 1994; Crissman et al., 1998). Data were collected on pesticide use, demographics and ailments that might have been related to pesticide use by pesticide applicators. Medical doctors made the health assessments, and the costs of ailments were regressed on pesticide use and other variables so the cost of pesticide-related illness could be estimated. Models were also used to estimate the relationship between labor productivity and health problems related to pesticides. Another cost-based approach includes assessing the cost of repairing environmental damage, for example the cost of treating pesticide-affected groundwater to make it potable. Another is to measure the cost of avoiding exposure to the environmental risk. For example, Abdalla et al. (1992) calculated the extra cost of purchasing drinking water in order to avoid drinking contaminated groundwater. Such measures tend to underestimate total environmental impacts because they omit consumer surplus (satisfaction above price paid) and they only focus on avoiding human exposure. Contingent valuation uses survey methods to collect data on people’s stated willingness to pay (WTP) to receive a benefit or their willingness to accept compensation for a loss. In the context of pest management, respondents might be asked how much they would be willing to pay to reduce the risk of pesticides to various categories of HE assets. The WTP data could later be linked to pesticide use data to arrive at a value for a change in pesticide use. Higley & Wintersteen (1992), Mullen et al. (1997) Swinton et al. (1999) and Cuyno et al. (2001) provide examples of using CV for such an assessment. Contingent valuation is subject to potential biases due to the way the survey is designed or


2.4.3 Trade-off approaches to compare economic and non-economic impacts Trade-off approaches avoid the complications of monetary valuation of pesticide risk, but they sacrifice the potential to aggregate market and nonmarket benefits in a single monetary measure. Trade-off analysis generally begins by plotting a given practice in two dimensions. Figure 2.3 illustrates a profitability measure plotted against a measure of health or environmental hazard. Each point represents a specific IPM practice or treatment, plotted according to its profitability and HE hazard level. Because profitability is desirable but HE hazards are not, the desired direction of movement is toward the upper left (northwest) of the figure. Points A, B and D represent increasing levels of profitability accompanied by increasing environmental hazard. Moving from point A to point B, there is a trade-off between a gain in profitability and an increase in HE hazard. Compared to points A and B, point C increases HE hazard without the gain in profitability that could be had from a mixture of practices A and B. Hence, point C is “inefficient” in the sense that more profitability per unit of environmental hazard could be had by a mixture of A and B, which lie on the trade-off frontier. Trade-off analysis can be used in at least three ways. First, it can be used to identify practices 4

B Profitability

administered, or to the hypothetical nature of the questions, so it should be used with care. An experimental auction technique can be used to minimize hypothetical bias. In an experimental auction, people use real money to bid on health or environmental improvements, and then pay their bids at the end of the experiment (Brookshire & Coursey, 1987). Hedonic pricing attempts to infer willingness to pay for environmental amenities from the prices of other goods based on the characteristics of those goods. This approach was used by Beach & Carlson (1993), who took a data set of herbicide attributes and estimated the value of herbicide safety from the effect of safety attributes on the herbicide prices.



Fig. 2.3 Illustrative trade-off frontier of profitability versus a hazard to health or environment.

that are not efficient, because alternative practices could achieve both objectives more effectively. Hoag & Hornsby (1992) plotted cost versus groundwater hazard index for various pesticides, finding that most were inefficient at attaining the twin objectives of lower cost and reduced groundwater hazard.4 Second, moves along the frontier can be used to measure the implied cost to a profitmaximizing farmer of adopting a more costly IPM practice with reduced HE hazards. For example, moving from point D to B or from B to A in Fig. 2.3 would entail sacrificing some profitability for the sake of reducing HE hazards. Third, trade-off analysis can be used to inform policy debates about who should bear the costs and enjoy the benefits of changing IPM practices (Antle et al., 2003).

2.5 Choosing an impact assessment method within a budget A balanced economic assessment of IPM impacts can divided into five basic parts: (1) definition of the IPM measure, (2) measurement of IPM adoption, (3) estimation of individual marketbased economic effects, (4) aggregation to market effects and (5) estimation of health and environmental (non-market) impacts. These parts can be evaluated with different amounts of detail,

The analytical tool of Hoag & Hornsby (1992) has been generalized by Nofzinger et al. (2002).




Table 2.1

Information needs for basic and extended IPM assessment protocols

Key assessment steps

Basic protocol (extended protocol in italic)

Define IPM measure

Identify IPM practices with direct, measurable economic and Health and Environmental (HE) impacts. Group practices into levels of IPM adoption or assign points in an adoption scale. Secondary trend data (e.g. national statistics). Ask experts to estimate maximum adoption, years to maximum adoption, and whether practices will become obsolete. Fit logistic curve to predict adoption over time. Ask farmers to keep detailed records on input use and yields over defined period of time. Conduct a grower adoption survey based on random sample to be extrapolated to population. Include questions on pest pressure, reasons for adoption, cost and yield, price received, and farm and farmer characteristics. Analyze determinants of adoption. Construct budgets (partial or enterprise). Build multi-year net present value (NPV) analysis. Regress profitability and crop yield on IPM practices and other explanatory variables to measure IPM effect. Risk analysis of IPM practices using stochastic dominance. Optimal pest control practices or whole-farm IPM effects using math programming. Multiply per-ha change in costs and returns by total ha under IPM (assumes IPM does not affect prices). Subtract annual IPM public costs. Conduct economic surplus analysis (allows price effects from aggregate IPM adoption), based on supply shift. Calculate NPV and internal rate of return based on annual net benefits from economic surplus and total IPM research and outreach costs. Identify changes in pesticide use as a result of adopting IPM. Do by pesticide a.i., rate, times sprayed, when applied, method of application, form applied. Regress pesticide use and aggregate hazard on IPM practices and other explanatory variables to measure IPM effect. Assess pesticide risk to water quality using spatial model (e.g. GLEAMS, CINDEX, etc.) Assess human health risks based on published, survey recall or medical visit data. Assess other non-target pesticide effects (e.g. to arthropods, birds, etc.) Weight HE effects in a monetary or non-monetary index or conduct trade-off analysis between profitability and HE criteria.

Measure adoption

Estimate individual economic impacts

Estimate aggregate impacts

Estimate health and environmental impacts

depending on the available time, money and expertise. Table 2.1 combines the methods described in the previous sections into elements for basic and extended protocols for IPM impact evaluation. The basic level covers data and meth-

ods for conducting the simplest elements of impact evaluation, while the extended protocol items offer ways to address multiple farmer or consumer objectives, risk and non-market valuation.


2.6 Conclusions Economic impacts of IPM programs can be measured at the level of the individual management unit or at the market level. Economic and environmental or health outcomes can be measured with differing levels of detail depending on time and budget constraints. This chapter briefly summarizes the major approaches to IPM impact assessment, guiding the reader to the key references for details.

References Abdalla, C. W., Roach, B. A. & Epp D. J. (1992). Valuing environmental quality changes using averting expenditures: an application to groundwater contamination. Land Economics, 68(2), 163–169. Alston, J. M., Norton, G. W. & Pardey P. G. (1995). Science under Scarcity: Principles and Practice for Agricultural Research Evaluation and Priority Setting. Ithaca, NY: Cornell University Press. Antle, J. M. & Pingali, P. L. (1994). Pesticides, productivity and farmer health: a Philippine case study. American Journal of Agricultural Economics, 76, 418–430. Antle, J., Stoorvogel, J., Bowen, W., Crissman, C. & Yanggen, D. (2003). The tradeoff analysis approach: lessons from Ecuador and Peru. Quarterly Journal of International Agriculture, 42, 189–206. Beach, D. & Carlson, G. A. (1993). Hedonic analysis of herbicides: do user safety and water quality matter? American Journal of Agricultural Economics, 75, 612– 623. Benbrook, C. M., Sexson, D. L., Wyman, J. A. et al. (2002). Developing a pesticide risk assessment tool to monitor progress in reducing reliance on high-risk pesticides. American Journal of Potato Research, 79, 183–199. Brookshire, D. S. & Coursey, D. L. (1987). Measuring the value of a public good: an empirical comparison of elicitation procedures. American Economic Review, 77, 554–566. Champ, P. A., Boyle, K. J. & Brown, T. C. (eds.) (2003). A Primer on Nonmarket Valuation. Dordrecht, Netherlands: Kluwer. Centro Internacional de Mejoramiento de Ma´ız y Trigo (1988). From Agronomic Data to Farmer Recommendations: An Economics Training Manual, completely revised edn. Mexico City, Mexico: Centro Internacional de Mejoramiento de Ma´ız y Trigo (CIMMYT). Cousens, R. (1987). Theory and reality of weed control thresholds. Plant Protection Quarterly, 2, 13–20.

Crissman, C. C., Antle, J. M. & Capalbo, S. M. (1998). Economic, Environmental and Health Tradeoffs in Agriculture: Pesticides and the Sustainability of Andean Potato Production. Dordrecht, Netherlands: Kluwer. Cuyno, L. C. M., Norton, G. W. & Rola, A. (2001). Economic analysis of environmental benefits of integrated pest management: a Philippines case study. Agricultural Economics, 25, 227–234. Cuyno, L., Norton, G. W., Crissman, C. C. & Rola, A. (2005). Evaluating the health and environmental impacts of IPM. In Globalizing Integrated Pest Management, eds. G. W. Norton, E. A. Heinrichs, G. C. Luther & M. E. Irwin, pp. 245–262. Ames, IA: Blackwell. Feder, G., Murgai, R. & Quizon, J. B. (2004). Sending farmers back to school: the impact of farmer field schools in Indonesia. Review of Agricultural Economics, 26, 45–62. Fernandez-Cornejo, J. (1998). Environmental and economic consequences of technology adoption: IPM in viticulture. Agricultural Economics, 18, 145–155. Freeman, A. M., III (2003). The Measurement of Environmental and Resource Values: Theory and Methods, 2nd edn. Washington, DC: Resources for the Future. Greene, C. R., Kramer, R. A., Norton, G. W., Rajotte, E. G. & McPherson, R. M. (1985). An economic analysis of soybean integrated pest management. American Journal of Agricultural Economics, 67, 567–572. Gutierrez, A. P., Regev, U. & Shalit, H. (1979). An economic optimization model of pesticide resistance: alfalfa and Egyptian alfalfa weevil. Environmental Entomology, 8, 101–107. Higley, L. G. & Wintersteen, W. K. (1992). A novel approach to environmental risk assessment of pesticides as a basis for incorporating environmental costs into economic injury levels. American Entomologist, 38, 34–39. Hoag, D. L. & Hornsby, A. G. (1992). Coupling groundwater contamination with economic returns when applying farm pesticides. Journal of Environmental Quality, 21, 579–586. Kovach, J., Petzoldt, C., Degni, J. & Tette, J. (1992). A Method to Measure the Environmental Impact of Pesticides, New York Food and Life Sciences Bulletin No. 139. Geneva, NY: Cornell University. New York State Agricultural Experiment Station. Available at 5203/. Martin, M. A., Schreiber, M. M., Riepe, J. R. & Bahr, J. R. (1991). The economics of alternative tillage systems, crop rotations, and herbicide use on three representative east central Corn Belt farms. Weed Science, 39, 299–397.




McNamara, K. T., Wetzstein, M. E. & Douce, G. K. (1991). Factors affecting peanut producer adoption of integrated pest management. Review of Agricultural Economics, 13, 129–139. Moffitt, L. J., Tanagosh, L. K. & Baritelle, J. L. (1983). Incorporating risk in comparisons of alternative pest management methods. Environmental Entomology, 12, 1003– 1111. Mullen, J. D., Norton, G. W. & Reaves, D. W. (1997). Economic analysis of environmental benefits of integrated pest management. Journal of Agricultural and Applied Economics, 29, 243–253. Musser, W. N., Tew, B. V. & Epperson, J. E. (1981). An economic examination of an integrated pest management production system with a contrast between E-V and stochastic dominance analysis. Southern Journal of Agricultural Economics, 13, 119–124. Napit, K. B., Norton, G. W., Kazmierczak, R. F. Jr. & Rajotte, E. G. (1988). Economic impacts of extension integrated pest management programs in several states. Journal of Economic Entomology, 81, 251–256. Norton, G. & Mullen, J. (1994). Economic Evaluation of Integrated Pest Management Programs: A Literature Review, Virginia Cooperative Extension Publication No. 448– 120. Blacksburg, VA: Virginia Tech. Norton, G. W., Swinton, S. M, Riha, S. et al. (2001). Impact Assessment of Integrated Pest Management Programs. Blacksburg, VA: Department of Agricultural and Applied Economics, Virginia Tech. Norton, G. W., Moore, K., Quishpe, D. et al. (2005). Evaluating socio-economic impacts of IPM. In Globalizing Integrated Pest Management, eds. G. W. Norton, E. A. Heinrichs, G. C. Luther & M. E. Irwin, pp. 225–244. Ames, IA: Blackwell. Pedigo, L. P., Hutchins, S. H. & Higley, L. G. (1986). Economic injury levels in theory and practice. Annual Review of Entomology, 31, 341–368.

Penrose, L. J., Thwaite, W. G. & Bower, C. C. (1994). Rating index as a basis for decision making on pesticide use reduction for accreditation of fruit produced under integrated pest management. Crop Protection, 13, 146– 152. Pingali, P. L., Marquez, C. B. & Palis, F. G. (1994). Pesticides and Philippine rice farmer health: a medical and economic analysis. American Journal of Agricultural Economics, 76, 587–592. Swinton, S. M. & Day, E. (2003). Economics in the design, assessment, adoption, and policy analysis of IPM. In Integrated Pest Management: Current and Future Strategies, R-140, ed. K. R. Barker, pp. 196–206. Ames, IA: Council for Agricultural Science and Technology. Swinton, S. M., Owens, N. N. & van Ravenswaay, E. O. (1999). Health risk information to reduce water pollution. In Flexible Incentives for the Adoption of Environmental Technologies in Agriculture, eds. F. Casey, A. Schmitz, S. Swinton & D. Zilberman, pp. 263–271. Boston, MA: Kluwer. Swinton, S. M., Renner, K. A. & Kells, J. J. (2002). On-farm comparison of three postemergence weed management decision aids in Michigan. Weed Technology, 16, 691–698. Szmedra, P. I., Wetzstein, M. E. & McClendon, R. (1990). Economic threshold under risk: a case study of soybean production. Journal of Economic Entomology, 83, 641–646. Teague, M. L., Mapp, H. P. & Bernardo, D. J. (1995). Risk indices for economic and water quality tradeoffs: an application to Great Plains agriculture. Journal of Production Agriculture, 8, 405–415. Zacharias, T. P. & Grube, A. H. (1986). Integrated pest management strategies for approximately optimal control of corn rootworm and soybean cyst nematode. American Journal of Agricultural Economics, 68, 704–715.

Chapter 3

Economic decision rules for IPM Leon G. Higley and Robert K. D. Peterson The year 2009 marks the 50th anniversary of the elaboration of the economic injury level (EIL) and the economic threshold (ET) concepts by Stern et al. (1959). The EIL and ET are widely recognized as the most important concepts in IPM. Because IPM is posited on the premise that certain levels of pests and pest injury are tolerable, the EIL and ET represent a crucial underpinning for any theory of IPM. Given the centrality of economic decision rule concepts to IPM, it follows that every IPM program should be based on these concepts. But this clearly is not the case. Why? In this chapter, we will discuss the historical development of economic decision levels, current approaches, and limitations of the EIL and ET. Finally, we will argue that EILs should be incorporated much more into IPM programs and that EILs are central to the continued development of environmental and economic sustainability concepts so important to IPM.

3.1 Economic decision levels The concept of tolerating pest injury was not introduced by Stern et al. (1959); it was discussed at least as early as 1934 (Pierce, 1934) but does not seem to have been developed further until 1959 (Kogan, 1998). In response to failures in pest control in California because of insecticide resistance by pests and insecticide mortality of natural

enemies, Stern et al. (1959) first proposed the fundamental concepts of the EIL, the ET, economic damage and pest status. The EIL is often defined as the lowest population density of pests that will cause economic damage. Economic damage is defined as the amount of injury that will justify the cost of control. The ET is defined as the density of pests at which control measures should be taken to prevent the pest population from reaching the EIL. In our remaining discussion we look at the EIL and ET in detail. However, these are not the only approaches for pest management decision making – many other sorts of decision models are possible. However, the factors used in establishing the EIL and ET are the same as those that must be considered in any management model. Moreover, the EIL and ET are often incorporated into more complex management models. Thus, whether as stand-alone criteria, or as components of more sophisticated and complex models, the EIL and ET have central roles in pest management decision making.

3.1.1 The EIL The EIL may be the simplest concept in all of applied ecology, yet it and the related concept of the ET continue to be misunderstood even by many IPM researchers and practitioners. The EIL is a straightforward cost–benefit equation in which

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



the costs (losses associated with managing the pest) are balanced with the benefits (losses prevented by managing the pest). Despite its simplicity, it was not until 1972 that a formula for calculating the EIL was introduced (Stone & Pedigo, 1972); see Pedigo et al. (1986) and Higley & Pedigo (1996) for detailed discussions of the historical evolution of the EIL equation. The EIL actually represents a level of injury, not a density of pests. However, numbers of pests per unit area are often used as an index for injury because injury can be very difficult to sample and measure. Using pest numbers as an index, EILs may be expressed as “larvae/plant,” “beetles/sweep,” “grass weeds/m2 ” or “moths/ trap” (Peterson & Higley, 2002). The most frequently used equation to determine the EIL is: EIL = C ÷ VIDK


where C = management costs per production unit (e.g. $/ha), V = market value per production unit ($/kg), I = injury per pest equivalent, D = damage per unit injury (kg reduction/ha/injury unit) and K = proportional reduction in injury with management (Pedigo et al., 1986). Despite the simplicity of the EIL concept, there is considerable complexity in the biologic components of the equation (Peterson & Higley, 2002). This is because injury per pest (I) and yield loss per unit injury (D) can be very difficult and costly to determine (Higley & Peterson, 1996). It is important that these components are known so that pests can be managed effectively within an IPM program, but because of the difficulty of determining I and D separately, both components often are determined together as simply loss per pest (Higley & Peterson, 1996; Peterson & Higley, 2002). When insects injure a plant through direct tissue loss, I may represent insect consumption rates. In contrast, with sucking insects direct measures of loss of photosynthate or nutrients may be impossible or impractical. Also, with many sucking insects it is necessary to characterize injury as a combination of quantity and time. Thus, an EIL may be defined for aphid-days (the injury– time combination), rather than an aphid density (although the ET may still be defined as a density, requiring an additional conversion from the EIL).

Variation in estimates of the damage relationship (how injury influences yield) follow from factors altering yield, most especially nutrient and water availability. The damage relationship, D, is the most variable component of the EIL (Higley, 2001; Peterson & Hunt, 2003). Because injury– yield loss relationships can change most strikingly between wet and dry conditions, where sufficient data are available separate EILs for normal and drought conditions may be determined (Hammond & Pedigo, 1982; Ostlie & Pedigo, 1985; Haile, 1999). Determining the damage relationship is the most difficult and limiting aspect of EIL development, not only because of environmental variability in yield responses. Experimentally, injury– yield loss determinations require quantification of injury, treatments consisting of different levels of injury and replication over a minimum of two years. Additionally, to determine a curvilinear relationship, at least three (and preferably four or more) levels of injury are necessary (Hammond, 1996; Higley & Peterson, 1996; Higley, 2001). Despite these difficulties, our theoretical understanding of injury–yield loss relationships is sound and evolving. The general nature of injury and yield, the damage curve, was originally defined by Tammes (1961) and has been demonstrated empirically in many systems (see Higley & Peterson, 1996 for a review). Subsequently, Pedigo and co-workers (Pedigo et al., 1986; Higley et al., 1993; Higley & Pedigo, 1996) assigned physiological interpretations to various portions of the damage curve and related these to different types of injury (Fig. 3.1). One application of the physiological approach is through the development of injury–yield loss relationships and associated EILs driven by an understanding of how defoliation by insects reduces light interception in soybean (Higley, 1992; Hammond et al., 2000; Malone et al., 2002). The EIL value determines the injury level, most often in the form of pest density, at which the pest management cost equals the cost from yield loss if no management occurs. For example, if the EIL is 5 larvae/plant, then economic damage is occurring at 5 larvae/plant. Producers and other decision makers should not wait until the EIL has been reached because at that level economic loss


Fig. 3.1 The damage curve, the general relationship between yield and increasing injury, and specific regions of the curve. A given injury–yield relationship may include all or part of this general function.

is already occurring. This is a common misunderstanding about the EIL. Therefore, the decision to initiate management activities, such as pesticide application, must be made before the EIL is reached so that economic damage can be prevented. Indeed, prevention of economic damage is, for all intents and purposes, the sole rationale underlying the EIL concept.

3.1.2 The ET In many IPM programs, the decision to initiate management action is based on the ET (Pedigo et al., 1986; Pedigo, 1996). Thus, the ET is the most widely used decision tool in IPM and is sometimes called the “action threshold.” Although defined in terms of the pest density at which management action should be taken, the ET is actually an index for when to implement pest management activities. For example, if an EIL is 10 larvae/plant, then an ET may be 8 larvae/plant. Action would be taken when 8 larvae/plant are sampled, not because that density represents an economic loss, but rather because it provides a window of time to take action before the pest density or injury increases to produce an economic loss (Peterson & Higley, 2002). Although it is often difficult to determine EILs, it is unarguably more difficult to calculate ETs. This is because of the risks associated with predicting when a population will exceed the EIL and with the variability in time delays for management action (Pedigo et al., 1986; Peterson, 1996).

Pedigo et al. (1989) divided ETs into two categories: subjective and objective. Subjective ETs are not based on calculated EILs, but rather on human experience. Because of this, they are practically always static values (e.g. 2 larvae/sweep-net sample regardless of changes in market value of the commodity and control cost). These subjective thresholds have been termed “nominal thresholds” by Poston et al. (1983). Despite their limitations and questionable accuracy in many situations, subjective ETs remain the most commonly used thresholds. Objective ETs are based on calculated EILs and therefore are inherently dynamic. This type of threshold has been further divided into three categories: fixed, descriptive and dichotomous (Pedigo et al., 1989). Fixed ETs are set at some percentage of the EIL and change proportionally with it. An example of a fixed ET is 80% of the EIL value. Descriptive ETs are based on estimates of pest population growth and dynamics and rely on accurate sampling to determine if the population will be likely to exceed the EIL (Wilson, 1985; Ostlie & Pedigo, 1987). Dichotomous ETs are based on statistical procedures to classify a pest population as “economic” or “non-economic.” The time-sequential sampling technique is the bestknown example of a dichotomous ET (Pedigo & van Schaik, 1984). In defense of subjective ETs, some pest situations do not readily fit the EIL model, because the pest-to-damage relationship is obscure or complex, experimental determination of economic damage is difficult or impossible, or pest tolerances are extremely low. In these situations, various thresholds similar to an ET but not explicitly related to an EIL have been proposed. Often these are termed action thresholds (AT), although we find this term problematic because it has been used both as a synonym for ET and as a different form of indicator for management action. In some situations entirely different decision rules based on non-economic criteria have been developed. For example, in instances where pest injury causes cosmetic damage, aesthetic thresholds have been developed (Sadof & Raupp 1996), some of which are economically based and are analogous to conventional ETs. Another example of an alternative decision rule is the sensory threshold




Galvin et al. (2007) developed based on changes in wine taste associated with the presence of multicolored Asian lady beetle, Harmonia axyridis, in grapes. Because an ET represents a time to take action, it is best suited for use with a regular sampling program. In some formal sampling/decision making procedures, like sequential sampling, the ET is essential for establishing decision points. With less sophisticated procedures, the ET still provides a benchmark against which pest densities are assessed. Because the EIL and ET concepts depend upon the principle of preventing injury, early assessments are important. Because natural mortality can impact final densities, various means for incorporating pest survivorship information into ETs have been developed (e.g. Ostlie & Pedigo, 1987; Brown, 1997; Barrigossi et al., 2003). These ETs provide a greatly improved assessment of the potential injuriousness of a population, and recognize the potential importance of age-specific mortality.

3.2 Development of EILs and ETs Development of calculated EILs has increased dramatically since 1972, the year the first calculated EIL was presented by Stone & Pedigo. Today, there are hundreds of published articles on EILs and ETs. Peterson (1996) reviewed the development of economic decision levels from 1959 through 1993, primarily by examining the scientific literature. Economic decision levels have been determined for more than 40 commodities, including most of the world’s major food and fiber crops. Greater than 80% of those decision levels have been for insect pests, with about 10%, 6% and 4% on mites, weeds and plant pathogens, respectively. Within the arthropod pests, nearly 50% of EILs have been determined for lepidopteran species, 17% for homopterans and 14% for coleopterans (Peterson, 1996). Based on injury type, 50% of EILs are for defoliators, 29% for assimilate sappers, 11% for mesophyll feeders (selective leaf and fruit feeders) and 10% for turgor reducers (root and stem feeders). More EILs have been determined for cabbage looper (Trichoplusia ni) than any other species. EILs for two-spotted spider mite (Tetrany-

chus urticae) have been determined for apple, common bean, cotton, grape and strawberry. Weed species for which EILs have been determined include Avena fatua, Datura stramonium, Helianthus annuus, Abutilon theophrasti, Amaranthus tuberculatus and Xanthium strumarium (Mortensen & Coble, 1996; Peterson, 1996). Why are there so many more decision levels for arthropods compared to weeds and plant pathogens? The EIL and ET were first defined and used by entomologists primarily because of the emerging need for insecticide resistance management and conservation of natural enemies. Also, arthropods were, and still are, largely amenable to curative management techniques, making economic decision levels for them relatively easy to incorporate into IPM programs. Plant diseases and, to some extent, weeds traditionally have been managed preventatively, largely precluding the need for economic decision levels. The rapid and widespread adoption of herbicide-resistant transgenic crop varieties has dramatically moved fieldcrop weed management to an ever greater reliance on preventative herbicide use, further reducing the need for curative decision tools, like EILs.

3.3 Current approaches Because of its applicability to many situations, advances in the EIL concept have occurred primarily through extensions of the model advanced by Pedigo et al. (1986). Aesthetic injury levels (AILs) have been determined based on attributes not readily definable in economic terms, such as form, color, texture and beauty. Examples of resources in which aesthetic injury levels could be used include lawns, ornamental plants, homes and public buildings. To ascertain value of these attributes, researchers have obtained input from owners or the general public using techniques such as the contingent valuation method (Sadof & Raupp, 1996). Using the contingent valuation approach, the damage per unit of pest injury has been determined in economic terms. Most results have revealed that public tolerance of pest injury is low, resulting in low AILs with acceptable levels of injury at or less than 10% (Raupp et al., 1988).


But, the salient point is that thresholds can be developed based on aesthetic considerations. In the early 1990s, environmentally based EILs were developed. Each variable in the EIL equation reflects management activities that could be manipulated to potentially enhance environmental sustainability. In particular, researchers have suggested incorporating environmental costs into the management cost variable, C (Higley & Wintersteen, 1992). The resulting EILs have been termed environmental economic injury levels (EEILs). Although the notion of reflecting environmental costs in farmer-level decision making received much interest, particularly among economists (e.g. Lohr et al., 1999; Brethour & Weersink, 2001; Florax et al., 2005), EEILs have not been implemented. Work by agricultural economists and others (e.g. Kovach et al., 1992) suggest that the value of the EEIL is in providing relative risk information to users and providing an economic context for comparing management options, rather than as a use/no-use criterion (like conventional EILs). Important conceptual advances in the ET occurred in the 1980s. An important theoretical and practical advance has been the conversion of insect population estimates into insectinjury equivalents. An insect-injury equivalent is the total injury potential of an individual pest if it were to survive through all injurious life stages (Ostlie & Pedigo, 1987). Injury equivalents are determined from estimates of pest population structure, pest density and injury potential. Incorporation of insect larval survivorship has led to a further refinement of the injury equivalency concept. Probably the most challenging goal for IPM is the establishment of multiple-species EILs. These EILs represent a potentially significant advance in IPM because they can provide decision makers with the ability to manage a complex of pests instead of managing single pest species (Peterson & Higley, 2002). The primary advances in this area have involved integrating pest injury from different species by determining if the multiplespecies injury has similar effects on the host. If different species produce injuries resulting in similar physiological responses by the host, then the pest species can be grouped into injury guilds (see

Welter, 1989; Higley et al., 1993; Peterson & Higley, 1993; Peterson & Higley, 2001 for more information on insect injury guilds). The injury guilds can then be used to characterize damage functions. One approach to determine multiple-species EILs has been to combine the injury guild concept with the injury equivalency concept (Hutchins et al., 1988). To develop multiple-species EILs using this concept, pest species must: (1) produce a similar type of injury, (2) produce injury within the same physiological time-frame of the host, (3) produce injury of a similar intensity and (4) affect the same plant part (Hutchins & Funderburk, 1991). A recent development in EIL theory is the probabilistic EIL (PEIL). The EIL typically is calculated by taking mean values for the parameters C, V, I and D. The K value often is set to one to indicate 100% efficacy for the control tactic. Despite the recognition that the EIL is determined by dynamic biological and economic parameters (Pedigo et al., 1986; Peterson, 1996), which can be highly variable and uncertain, there has been little effort to quantify uncertainty and to use estimates of uncertainty in the determination of EILs (Peterson, 1996). Therefore, Peterson & Hunt (2003) defined the PEIL as an EIL that reflects its probability of occurrence. The probability of occurrence is determined by incorporating the variability and uncertainty associated with the input variables used to calculate the EIL. Peterson & Hunt (2003) used Monte Carlo simulation, a random sampling technique in which each input variable in the model was sampled repeatedly from a range of possible values based on each variable’s probability distribution. Then, the variability for each input was propagated into the output of the model so that the model output reflected the probability of values that could occur. How, then, can PEILs be used? The practical value of the PEIL is that multiple EIL values ranked as percentiles as a result of the Monte Carlo distributional analysis allow the decision maker to choose her or his tolerable level of risk within an IPM program (Peterson & Hunt, 2003). For example, if the decision maker is risk averse (i.e. she does not want to risk economic damage even if it means spraying in the absence of economic damage) and needs to decide which threshold to use for bean leaf beetle (Cerotoma trifurcata)




in seedling-stage soybean, she may choose a PEIL of 7.7 adults/plant. This PEIL represents the 25th percentile of values as determined from the model output. The use of 7.7 adults/plant as the PEIL ensures that an EIL 5 sprays) Mean Simple average a b

Action to choose Probabilityb

Biologically based IPM

Conventional grower system

0.70 0.20 0.10

1795 775 1381 1550 1317

285 −366 871 213 263

Hutchison et al. (2006a, b); Burkness & Hutchison (2008). Subjective probabilities based on ten-year experience of IPM specialist (E. Burkness) with cabbage.

experience; Box 4.1). With numerical outcomes and probabilities, the situation is risky in the sense that all outcomes and probabilities are known, including how the decision maker’s actions affect the outcomes. This simplistic payoff matrix (or an equivalent decision tree or statistical function) could be constructed to represent one of the many existing EIL-based IPM programs, with the actions, events and outcomes expanded to better match real situations (Mumford & Norton, 1984). However, we want to broaden the conceptual scope to illustrate how these same representations can be used to describe and analyze risk and IPM from a more strategic perspective, as opposed to the more common tactical/operational approach of EIL-based IPM. Hence we examine management alternatives (biologically based IPM versus a conventional grower system) for cabbage looper (Trichoplusia ni) in cabbage (Brassica oleracea), based on a Minnesota cabbage IPM case study of Hutchison et al. (2006a, b) and Burkness & Hutchison (2008). Table 4.2 is a payoff matrix representation for the case study. The action is the grower’s pest management system (IPM or conventional) and events are the pest intensity measured by the number of sprays needed with IPM. Outcomes (net returns $/ha) are based on observed yields and prices, while event probabilities are based on the ten-year experience of a Minnesota

IPM specialist (E. Burkness) working with cabbage (Hutchison et al., 2006b, Burkness & Hutchison, 2008). For the same case study, Fig. 4.1 is the equivalent decision tree representation. Figure 4.2 is the probability density function and cumulative distribution function for the random net return outcomes. Commonly, these functions are smooth curves such as the normal (bell) curve, but here, with only three outcomes, Fig. 4.2 is the result. In later discussion, we develop a representation of the case study with more outcomes, which is cumbersome to represent as a payoff matrix or decision tree, but gives smoother functions. The payoff matrix and decision tree representations show that most of the time, pest pressure is low, which implies low net returns for the conventional (scheduled) system because of unneeded insecticide applications, but high returns with IPM because of the cost savings from not applying unneeded sprays. Moderate pest pressure in about one in five years implies higher returns with IPM than with the conventional system, primarily because consistent scouting reliably detects pest intensity and uses timely control, whereas the conventional grower misses these events often enough either to incur significant losses or to apply too late or when not needed. Finally, high pest pressure in about one in ten years implies higher returns for both cases, but IPM outperforms the conventional system because it more



Probability Density

Fig. 4.1 Decision tree representation of net returns ($US/ha) for the Minnesota 1998–2001 cabbage IPM case study using subjective probabilities based on ten-year experience of IPM specialist (E. Burkness) (Hutchison et al., 2006b; Burkness & Hutchison, 2008).

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -500

IPM Conventional






Fig. 4.2 Probability density function and cumulative distribution function for the Minnesota 1998–2001 cabbage IPM case study using subjective probabilities based on ten-year experience of IPM specialist (E. Burkness) working in cabbage pest management (Hutchison et al., 2006a, b; Burkness & Hutchison, 2008).

Net Returns ($/ha)

Cumulative Probability


1.0 0.8 0.6

IPM Conventional

0.4 0.2 0.0 -500






Net Returns ($/ha)

closely tracks the pest pressure and actually uses more applications, since they are justified. Figure 4.2 shows the probability density function for the conventional system is shifted left because low outcomes are more likely, while shifted right for IPM because high outcomes are more likely. These shifts imply that the cumulative distribution function for the conventional system rises more quickly to 1.0 and is everywhere to the left of the cumulative distribution function for IPM.

For decision making, however, more specific measures of the value and/or riskiness of these two pest management systems seem warranted, as these qualitative descriptions are somewhat unsatisfactory. How does one say a certain case is more or less risky than another? Several measures exist that allow quantitative comparison of risky cases. First are measures of central tendency or location that indicate where the risky outcome tends to result. Most common among these is the statistical mean or expected


Risk measures for net returns ($US/ha) for the Minnesota 1998–2001 cabbage IPM case studya with both subjective and objective probabilitiesb

Table 4.3

Subjective probabilities Measure Probabilities unknown Minimum Maximum Simple average Simple variance Simple standard deviation Probabilities known Mean Median Mode Variance Standard deviation Coefficient of variation Return–risk ratioc Break-even probability Probability of $1000/ha Certainty equivalents Mean–Variance (β = 0.0005) Constant absolute risk aversion (CARA) (r = 0.0001) Constant relative risk aversion (CRRA) (ρ = 1.2)

Biologically based IPM

Objective probabilities

Conventional system

Biologically based IPM

775 1 795 1 317 263 172 513

−366 871 263 382 894 619

1 550 1 795 1 795 164 998 406 0.26 3.81 1.0 0.8

213 285 285 113 973 338 1.58 0.63 0.8 0.0

1 177 1 560 1 896 940 522 970 0.82 1.21 0.88 0.75

1467 1541

156 208

706 1127


Conventional system

−1 411 1 949 1 007 1 049 102 1024

−3 998 2 050 −128 3 656 829 1912 186 771 1 107 & 1272 2 919 525 1 709 9.20 0.11 0.71 0.50 −1274 28 –


Hutchison et al. (2006a, b); Burkness & Hutchison (2008). Subjective probabilities based on ten-year experience of IPM specialist (E. Burkness) working in cabbage pest management. Objective probabilities based on four-year data set of Hutchison et al. (2006a, b); Burkness & Hutchison (2008). c Return–risk ratio is similar in concept to the Sharpe ratio (Sharpe, 1994), but in this case the ratio is based on the probabilistic mean and standard deviation of net returns; values >1 are preferred (see text). b

value. For outcomes zi each with probability pi ,  the definition of the mean is µz = i pi zi (Boehlje & Eidman, 1984; Freund, 1992). The mean is sometimes called the probability weighted average of the outcomes or the probabilistic mean to distinguish it from the simple average of the outcomes:  ¯z = n1 i zi . The first two columns of Table 4.3 report these values for both systems in the case study for the subjective probabilities of the three outcomes. The means imply that on average, IPM

generates greater net returns than the conventional system ($US 1550/ha versus $US 213/ha). These means do not equal the simple average of the three outcomes, since the probabilities of each outcome are not equal. Two other common measures of central tendency are the median and the mode (Freund, 1992). With symmetric distributions, the median and mode will be similar in magnitude to the mean; a large difference between these and the




mean indicates a skewed distribution. The median is the middle outcome – the outcome with half the outcomes above and half below. When outcome probabilities are unequal (the case here), the median is the value on the horizontal axis where the cumulative distribution function equals 0.50 (also called the 50th percentile). As shown in Table 4.3, the median with IPM is $US 1795/ha and $US 285/ha for the conventional system. The mode is the most likely outcome, which here is the same as the median (Table 4.3). For risky outcomes, measures of dispersion are used to provide some indication of the variability of outcomes. Most common are the variance and standard deviation, though the coefficient of variation and return–risk ratio are useful as well. For outcomes zi each with probability pi  and mean µz , the variance is σz2 = i pi (zi − µz )2 and the standard deviation is the positive square root of the variance (Boehlje & Eidman, 1984; Freund, 1992). The standard deviation has the same units as the mean and can be approximately interpreted as the probability weighted average deviation of outcomes from the mean. The variance is sometimes called the probabilistic variance to distinguish it from the simple variance of out comes: n1 i (zi − ¯z)2 , with the comparable distinction made for the probabilistic standard deviation. Table 4.3 reports these measures, showing that the standard deviation of net returns is greater with IPM than with the conventional system ($US 406/ha versus $US 338/ha), implying that net returns with IPM are “riskier” since they are more variable or dispersed. However, some would argue against this interpretation since mean net returns are larger with IPM as well, so the greater standard deviation is relatively less important. Hence, the coefficient of variation (CV) and the risk–return ratio are measures of risk that normalize for differences in the mean. They are unit-less measures and thus useful for comparing risk-adjusted returns for alternative IPM systems. The CV is the standard deviation divided by the mean (σz /µz ) and the risk–return ratio adapted to the IPM context is the mean divided by the standard deviation (µz /σz ). The CVs in Table 4.3 indicate that IPM is less risky since the standard deviation is much smaller relative to mean returns (0.26 versus 1.58). The risk–return ratio is used

in finance to measure the expected return per unit of risk, where expected returns are net of the risk-free return (e.g. the bond market) and risk is measured by the volatility (the standard deviation of the change in returns) (Sharpe, 1994). In the context of IPM, no equivalent of the riskfree return exists, so the mean return is used, and the standard deviation of returns replaces the volatility, as the change in net returns is usually not tracked continuously. In Table 4.3 the risk–return ratio also indicates that IPM is less risky than the conventional system, since IPM has greater return for the associated risk (3.81 versus 0.63). A problem with these measures is that they treat risk symmetrically – variability from higher than expected returns is treated the same as variability from lower than expected returns. In many contexts, including IPM, more specific risk measures are useful to differentiate between higher than expected returns (upside risk) and lower than expected returns (downside risk). Statistical measures of skewness describe asymmetric probability distributions (Freund, 1992), but are little used in IPM. However, the Value at Risk (VaR) is a more practical measure of risk borrowed from finance to address these issues (Manfredo & Leuthold, 1999). Value at Risk focuses on downside risk (the occurrence of lower than expected returns) by reporting the returns associated with a chosen critical probability; specifically, with critical probability α c and cumulative distribution function F(z) for outcomes z, VaR(α c ) = F −1 (αc ). Figure 4.3 (plot A) illustrates the derivation of the VaR. For example, if the VaR(5%) is $US 50/ha, then net returns will be $US 50/ha or less with 5% probability. Here, the larger the VaR for a given probability the less risky the alternative, which is opposite from the original finance context for portfolio values. Also, note that the median is the VaR for a critical probability of 0.50. An issue with using VaR is what probability to use. In finance, 5% and 1% are common (Manfredo & Leuthold, 1999), but in agriculture and IPM, the appropriate probability is not clear. For example, how useful is it to know that net returns will be $US 50/ha or less with 5% probability? A closely related measure that seems more


Fig. 4.3 Derivation of Value at Risk (VaR) for a given critical probability α c (plot A) and of the probability of equaling or exceeding the target outcome zt = Pr(z ≥ zt ) (plot B), both in terms of the cumulative distribution function F(z) for outcomes z.

appropriate for crop production is to reverse the VaR process and report the probability for a critical outcome (rather than the VaR’s outcome for a critical probability). A simple example is the breakeven probability – the probability that net returns will be zero or better (e.g. Mitchell, 2005). However, using zero for the critical outcome is arbitrary; decision makers may be interested in the probability of achieving other target outcomes or profit goals. This probability, the probability of equaling or exceeding the target outcome zt , is defined in terms of the cumulative distribution function as Pr(z ≥ zt ) = 1 − F (zt ). Figure 4.3 (plot B) illustrates the derivation of this probability. For example, if the probability of equaling or exceeding the target outcome of $US 100/ha is 0.10, then in one out of ten years, the grower will earn at least $US 100/ha. For a given target outcome, the larger the probability of equaling or exceeding the target outcome, the lower the risk – a 0.1 probability of equaling or exceeding $100/ha is less risky than a 0.05 probability.

Deriving the VaR and probability of achieving target outcomes for the current specification of the cabbage IPM case study is less useful because with only three outcomes, the cumulative distribution functions have long vertical and horizontal sections implying constant VaR for wide ranges of critical probabilities and constant probabilities for wide ranges of target outcomes (Fig. 4.2). Nevertheless, Table 4.3 reports the break-even probability and the probability of returns reaching at least $US 1000/ha for the case study. To smooth these curves, we expand the number of outcomes using experimental data (Hutchison et al., 2006b; Burkness & Hutchison, 2008) to switch from subjective probabilities to objective probabilities. Subjective probabilities are useful when few formal data exist for the situation other than personal experience. In such cases, having only a few events as in the case study is reasonable, but implies functions as plotted in Fig. 4.2. However, in research contexts, more formal data can be collected and analyzed, giving smoother functions with objective probabilities. For this case study, data from 24 observations (six fields per year × four years) of net returns for each system give smoother functions to use to illustrate these measures of risk (Hutchison et al., 2006b; Burkness & Hutchison, 2008). Figure 4.4 shows the traditional histograms of net returns (plotting the probability as points as in Fig. 4.2 is difficult to interpret) and the cumulative distribution functions. The last two columns in Table 4.3 report all previously discussed measures of risk, including the break-even probability and probability of returns reaching at least $US 1000/ha. Figure 4.4 shows a negative skew for the returns of both density functions, which drives the large difference between the means and the medians/modes for both systems (Table 4.3). Interestingly, the conventional system is now “riskier” than IPM based on both absolute measures of risk (variance, standard deviation) and relative measures (CV, return–risk ratio), as well as the two probability-based measures (break-even probability, probability of achieving at least $US 1000/ha). This use of more outcomes and smoother functions can continue until the density and distribution functions are continuous plots similar to




9 8 7 6 5 4 3 2 1 0 -4000 -3000 -2000 -1000

IPM Conventional



Figure 4.4 Histogram and cumulative distribution function for the Minnesota 1998–2001 cabbage IPM case study using objective probabilities based on 24 observations for each system (Hutchison et al., 2006b; Burkness & Hutchison, 2008).


Net Returns ($/ha)

Cumulative Probability


1.0 0.8 0.6

IPM Conventional

0.4 0.2 0.0 -4000 -3000 -2000 -1000




Net Returns ($/ha)

Fig. 4.3. The only major change is that calculating the risk measures in Table 4.3 requires use of different formulas than given previously (integration rather than summation). For common distributions, texts give formulas for the mean, standard deviation, etc. (e.g. Evans et al., 2000) and software packages give numerical solutions for values of the cumulative distribution and/or its inverse when closed form expressions do not exist, such as for the normal distribution.

4.5 Decision making criteria and tools These representations of risky pest management situations and the associated measures of the risk under different actions still leave unanswered the question of which action to choose. Combining a representation of the risky situation with a decision criterion gives a decision making tool, a method to identify which actions are optimal given the specific criteria. Clearly, the optimal action depends on the criterion. Here we describe

criteria commonly used in agriculture and pest management. The criteria can first be separated into those that require probabilities and those that do not (Boehlje & Eidman, 1984). Maximin (also called minimax), a criterion adapted from game theory not using probabilities, identifies the worst outcome for each action and chooses the action with the best outcome among these worst-case scenarios. The names arise because the criterion maximizes the minimum outcome (or minimizes the maximum loss), a fairly conservative approach (Boehlje & Eidman, 1984). Maximax, a related criterion little used in practical applications, but useful as an overly optimistic benchmark, identifies the best outcome for each action and chooses the action with the best outcome among these best-case scenarios. Finally, another criterion ignores all probability assessments and chooses the action with the greatest simple average for its outcomes. In some sense, this criterion is comparable to an uninformed prior in Bayesian statistics – initially all outcomes are equally possible (Boehlje & Eidman, 1984). Table 4.3 shows that IPM is optimal with both the maximin and the maximax criteria,


whether using the three-state or 24-state example, the minimum (and maximum) outcome for IPM exceeds the minimum (and maximum) outcome for the conventional system. The simple average of returns is also greater for IPM. We are unaware of published examples using these criteria for IPM. The most common criterion using probabilities optimizes the expected (mean) outcome, for example choosing the action with the greatest mean net return (as explained below, this decision maker is called “risk neutral”). For our case study, Table 4.3 shows that IPM is the optimal action under this criterion – mean returns with IPM exceed mean returns with the conventional system using subjective or objective probabilities. Applications of this criterion in pest management, with or without assumed probabilities, are often implied in IPM applications, but the probabilities are not typically defined (e.g. Mumford & Norton, 1984; Burkness & Hutchison, 2008). Various safety-first criteria exist, for example minimizing the probability that returns fall below a critical level. This criteria places no value on mean returns, but other safety-first criteria do, such as maximizing mean returns, subject to ensuring a specific probability that a certain minimum return is achieved (e.g. choosing the action with greatest mean returns subject to maintaining a 5% probability that returns are at least $US100/ha). Safety-first criteria are useful for IPM with limited income farmers such as in developing countries where a certain income level is needed for survival (e.g. Norton et al., 2005). Maximizing mean returns (or expected profit) ignores variability – actions with the same mean returns are equal under the criterion, even if their variability differs. However, most people are willing to trade off between the mean and the variability of returns, preferring actions with lower mean returns because they are less variable. For example, many people voluntarily buy insurance because the indemnity reduces the variability of returns, though the premium reduces their mean return. How people trade between certain and variable outcomes and between the mean and vari-


ability of returns is an active area of empirical and theoretical research in many fields beyond the scope of this chapter (e.g. Camerer, 2003; Eeckhoudt et al., 2005; Ernst & Paulus, 2005; Fecteau et al., 2007; Mitchell & Onstad, 2008). Here we describe and illustrate some of the more common methods used in agricultural contexts.

4.5.1 Risk preferences defined A fundamental assumption of these methods is that individuals implicitly impose a cost on mean returns to adjust for their variability, similar to a discount factor imposing a cost on future returns. “Risk preferences” is the term used in economics and related disciplines to describe how individuals trade between risk and returns (see Box 4.1). In many contexts, people are risk averse – willing to give up some returns to obtain reduced variability, for example by buying insurance. However, in other contexts, people will give up mean returns to obtain increased variability, for example by buying a lottery ticket. In the first case, the person is called “risk averse,” and “risk taking” in the second case (Chavas, 2004; see Box 4.1).1 Depending on the context, the same person can exhibit both behaviors (e.g. buy insurance and lottery tickets). Two concepts are often used to understand this trading between risk and returns. For a person facing a risky outcome, the certainty equivalent is the non-random payment that makes the person indifferent between taking the risky outcome and this certain payment. The risk premium is the difference between the expected outcome and this certainty equivalent. For example, if a risky outcome has a mean return of $US 500 and a person’s certainty equivalent is $US 450, then this person’s risk premium is $US 50. If the certainty equivalent is less than the mean return (or the risk premium is positive), the person is risk averse. Usually this trading between risk and returns is described using a preference (or utility) function to convert risky outcomes into a measure describing the benefit after accounting for risk; a common example is mean–variance preferences.

Unfortunately, this terminology is not fully standardized. Many use “risk loving” for “risk taking,” while Hardaker et al. 2004 use “risk preference” for “risk loving” and “risk attitudes” when most use “risk preferences” (Chavas, 2004; Eeckhoudt et al., 2005).




For a random outcome z, a person’s outcome is u(z) = µz − βσz2 , where µz and σz2 are the respective mean and variance of z, β is a parameter describing how the individual adjusts the mean for risk as measured by the variance, and u(·) is called the utility function. For this case, u(z) is the certainty equivalent and βσz2 is the risk premium. If β > 0, the person is risk averse, if β < 0, the person is risk loving. If β = 0, then u(z) = µz ) and the person is “risk neutral” – outcome variability does not matter, only the mean return. Within risk averse preferences, further subtypes of risk preferences are defined. Two of the most common are constant absolute risk aversion (CARA) and constant relative risk aversion (CRRA). CARA implies that the risk premium is independent of the initial wealth level. However, the risk premium with CARA depends on the units of measure (e.g. € versus $, or $/acre versus $/ha) (Chavas, 2004). Hence, as a convenient normalization, the relative risk premium expresses the risk premium as a percentage of expected final wealth. CRRA implies that this relative risk premium is constant with respect to wealth, though the absolute risk premium is not. However, the general functional form implying CRRA is undefined for negative outcomes, limiting its use for some applications without adjustments. Mathematically, the CARA utility function is u(z) = −exp(−r z), where r > 0 is the coefficient of absolute risk aversion, while the CRRA utility function is u(z) = z1−ρ when ρ > 1, u(z) = −z1−ρ when ρ < 1 and ln(z) when ρ = 1, where ρ is the coefficient of relative risk version (Chavas, 2004). Finally, note that concavity of the utility function u(z) over the range of outcomes z implies risk aversion (i.e. a positive risk premium, or a certainty equivalent less than the expected outcome µz ). These descriptions are brief and far from comprehensive, plus all presented cases follow the von Neumann–Morganstern expected utility model, which has well-known limits for accurately describing risky decisions (Chavas, 2004). Other types of utility functions based on alternative decision criteria or theories exist, including rank-dependent expected utility, ambiguity


aversion, and loss aversion/prospect theory. See Chavas (2004), Eeckhoudt et al. (2005) and Hardaker et al. (2004) for more comprehensive and detailed descriptions and examples of these. Table 4.3 summarizes certainty equivalents for mean–variance, CARA and CRRA preferences (when possible) for both systems with both sets of probabilities, with specific parameter values used for utility functions reported in parentheses. These parameters were chosen purely for illustration; different parameters would generate different results. Babcock et al. (1993) and McCarl & Bessler (1989) provide guidance on parameter choices. IPM is optimal for both sets of probabilities with these parameters. This is not surprising, given the notably larger mean and lower variability with IPM. For these data, it would be difficult to identify a reasonable model of risk averse preferences that made the conventional pest management approach optimal. Applications of these expected utility criteria in pest management include Hurley et al. (2004), Mitchell et al. (2004) and Mitchell (2008). Stochastic dominance is an alternative criterion that is theoretically attractive as it imposes little structure on preferences (i.e. no utility function is required) other than positive slope and possibly concavity. Stochastic dominance ranks the various actions a decision maker has by comparing the cumulative distribution functions of outcomes that result with each action. Without explaining the theoretical foundations (see Boehlje & Eidman, 1984; Chavas, 2004; Hardaker et al., 2004), first-order stochastic dominance (FOSD)2 implies that, if the cumulative distribution with action A is everywhere less than or equal to the cumulative distribution with action B, then action A is preferred to action B by all decision makers with positively sloped utility functions (i.e. people who prefer more income to less). Alternatively, A first order stochastically dominates B if the entire cumulative distribution function for A is to the right of the function for B. Mathematically, FOSD requires FA (z) ≤ FB (z) for all outcomes z, where Fi (·) is the cumulative distribution for i ∈ {A, B}. Figure 4.2 shows that IPM first order

Some use the terminology first degree stochastic dominance (Hardaker et al., 2004).


stochastically dominates the conventional system for the case study with subjective probabilities. The implication is that almost everyone would prefer IPM to the conventional system in this case. Second-order stochastic dominance (SOSD) is defined in terms of the area between the cumulative distribution for actions A and B. Specifi z∗ cally, if zmin F B (z) − F A (z)dz ≥ 0 for all zmin ≤ z ∗ ≤ zmax , where zmin and zmax are the minimum and maximum possible values for z, then action A second order stochastically dominates action B so that all decision makers with concave utility functions (i.e. who are risk averse) prefer action A to action B. SOSD requires that the area between the cumulative distributions (above A and below B) be positive, beginning with the minimum outcome zmin and up to any outcome z∗ , for all possible z∗ . Figure 4.4 shows that IPM second order stochastically dominates the conventional system for the case study with objective probabilities because the cumulative distributions cross only once, FA (·) is initially above FB (·), and the area when FB (·) is above FA (·) is much larger than the area when FA (·) is above FB (·). The implication is that all risk averse individuals would prefer IPM to the conventional system in this case. Finally, note that stochastic dominance is only a partial ordering. Cumulative distributions can occur that cannot be ordered by FOSD or SOSD (Chavas, 2004; Hardaker et al., 2004). Also, Hardaker et al. (2004) discuss additional types of stochastic dominance. Besides the brief examples provided here by the case study, previous applications of the expected value (mean) and variance (or standard deviation) in IPM include: plant pathogens (Carlson, 1970; Norton, 1976), arthropods (Burkness et al., 2002) and alternative weed management systems (Hoverstad et al., 2004). Stochastic dominance has been used to assess alternative IPM tactics or programs for arthropods (Moffitt et al., 1983; Burkness et al., 2002), weed management (Hoverstad et al., 2004) and multiple pest IPM (Musser et al., 1981; see review by Fox et al., 1991).

represents just one of several “decision tools” or approaches to decision making. Other tools compare the economic performance, or risk adjusted economic returns for IPM versus other pest management systems (e.g. organic or conventional). The EIL is most often used as a therapeutic (ex post) decision criterion for managing pest populations during a given growing season. We have summarized alternative decision making approaches that are applicable to IPM that can be used with the appropriate information to assess the value and risk of IPM in advance of a growing season (ex ante). As such, these methods facilitate objective comparison of IPM performance with alternative production systems. Most of the methods reviewed also incorporate some measure of economic risk, such as the variance or standard deviation of net returns, to facilitate decision making under uncertainty. More work is needed to examine the economic impact of multiple arthropods, weeds and/or pathogens on the value and risk of IPM for a given crop, and for two or more crops at the farm or enterprise level (see Fox et al., 1991; Chapter 2). As seed treatments and biotechnology become increasingly popular (e.g. maize growers in the USA), we anticipate that these methods may become more common in future ex ante assessments of IPM. In addition, similar approaches could be taken to better understand the value and risk of environmental impacts of IPM versus alternative systems. The increasing use of environmental indices to evaluate pest management programs has been beneficial (e.g. Kovach et al., 1992; Benbrook et al., 2002). However, more studies like those of Mullen et al. (1997) and Edson et al. (2003), and expanded methods for environmental analysis (e.g. Shiferaw et al., 2004; Jepson, 2007), are needed so growers and policy makers can fully assess both the value and risk of economic and environmental outcomes for IPM and competing pest management systems.


4.6 Conclusions In summary, we have shown that although the EIL is foundational to successful IPM programs, it

Babcock, B. A., Choi, E. K. & Feinerman, E. (1993). Risk and probability premiums for CARA utility functions. Journal of Agricultural and Resource Economics, 18, 17–24. Bacic, I. L. Z., Bregt, A. K. & Rositer, D. G. (2006). A participatory approach for integrating risk assessment into




rural decision-making: a case study in Santa Catarina, Brazil. Agricultural Systems, 87, 229–244. Barry, P. (ed.) (1984). Risk Management in Agriculture. Ames, IA: Iowa State University Press. Bechinski, E. J. (1994). Designing and delivering in-field scouting programs. In Handbook of Sampling Methods for Arthropods in Agriculture, eds. L. P. Pedigo & B. D. Buntin, pp. 683–706. Boca Raton, FL: CRC Press. Benbrook, C. M., Sexson, D. L., Wyman, J. A. et al. (2002). Developing a pesticide risk assessment tool to monitor progress in reducing reliance on high-risk pesticides. American Journal of Potato Research, 79, 183– 200. Boehlje, M. D. & Eidman, V. R. (1984). Farm Management. New York: John Wiley. Burkness, E. C. & Hutchison, W. D. (2008). Implementing reduced-risk IPM in fresh-market cabbage: Improved net returns via scouting and timing of effective management. Journal of Economic Entomology, 101, 461– 471. Burkness, E. C., Hutchison, W. D., Weinzierl, R. A. et al. (2002). Efficacy and risk efficiency of sweet corn hybrids expressing a Bacillus thuringiensis toxin for lepidopteran pest management in the Midwestern U.S. Crop Protection, 21, 157–169. Camerer, C. (2003). Behavioral Game Theory: Experiments in Strategic Interaction. Princeton, NJ: Princeton University Press. Carlson, G. A., (1970). The decision theoretic approach to crop disease prediction and control. American Journal of Agricultural Economics, 52, 216–223. Chavas, J. P. (2004). Risk Analysis in Theory and Practice. New York: Elsevier. Crowder, D. W., Onstad, D. W., Gray, M. E. et al. (2005). Economic analysis of dynamic management strategies utilizing transgenic corn for control of western corn rootworm (Coleoptera: Chrysomelidae). Journal of Economic Entomology, 98, 961–975. Cuperus, G. W. & Berberet, R. C. (1994). Training specialists in sampling procedures. In Handbook of Sampling Methods for Arthropods in Agriculture, eds. L. P. Pedigo & B. D. Buntin, pp. 669–681. Boca Raton, FL: CRC Press. Edson, C., Swinton, S., Nugent, J. et al. (2003). Cherry Orchard Floor Management: Opportunities to Improve Profit and Stewardship, MSU Ext. Bull. E-2890. East Lansing, MI: Michigan State University. Eeckhoudt, L., Gollier, C. & Schlesinger, H. (2005). Economic and Financial Decisions under Risk. Princeton, NJ: Princeton University Press. Ehler, L. E. & Bottrell, D. G., (2000). The illusion of integrated pest management. Issues in Science and Technology, 16, 61–64.

Ernst, M. & Paulus, M. P. (2005). Neurobiology of decision making: a selective review from neurocognitive and clinical perspective. Biological Psychiatry, 58, 597– 604. Evans, M., Hastings, N. & Peacock, B. (2000). Statistical Distributions, 3rd edn. New York: John Wiley. Feather, P. & Cooper, J. (1995). Voluntary Incentives for Reducing Agricultural Nonpoint Source Water Pollution, Agricultural Information Bulletin No. 716. Washington, DC: US Department of Agriculture, Economic Research Service. Fecteau, S., Pascual-Leone, A., Zald, D. H. et al. (2007). Activation of prefrontal cortex by transcranial direct current stimulation reduces appetite for risk during ambiguous decision making. Journal of Neuroscience, 27, 6212–6218. Feder, G. (1979). Pesticides, information, and pest management under uncertainty. American Journal of Agricultural Economics, 62, 97–103. Foster, R. E., Tollefson, J. J., Nyrop, J. P. & Hein, G. L. (1986). Value of adult corn rootworm (Coleoptera: Chrysomelidae) population estimates in pest management decision making. Journal of Economic Entomology, 79, 303– 310. Fox, G., Weersink, A., Sarwar, G., Duff, S. & Deen, B. (1991). Comparative economics of alternative agricultural production systems: a review. Northeastern Journal of Agricultural and Resource Economics, 20, 124–142. Freund, J. E. (1992). Mathematical Statistics, 5th edn. Englewood Cliffs, NJ: Prentice Hall. Gollier, C. (2001). The Economics of Risk and Time. Cambridge, MA: MIT Press. Grieshop, J. I., Zalom, F. G. & Miyao, G. (1988). Adoption and diffusion of Integrated Pest Management Innovations in Agriculture. Bulletin of the Entomological Society of America, 34, 77–78. Gutierrez, A. P. & Baumg¨ artner, J. (2007). Modeling the dynamics of tritrophic population interactions, In Perspectives in Ecological Theory and Integrated Pest Management, eds. M. Kogan & P. Jepson, pp. 301–360. Cambridge, UK: Cambridge University Press. Hammond, C. M., Luschel, E. C., Boerboom, C. M. & Nowak, P. J. (2006). Adoption of integrated pest management tactics by Wisconsin farmers. Weed Technology, 20, 756–767. Hammond, R. B. (1996). Limitations to EILs and thresholds, In Economic Thresholds for Integrated Pest Management, eds. L. G. Higley & L. P. Pedigo, pp. 58–73. Lincoln, NE: University of Nebraska Press. Hardaker, J. B., Huirne, B. R. M., Anderson, J. R. & Lien, G. (2004). Coping with Risk in Agriculture, 2nd edn. Wallingford, UK: CABI Publishing.


Hoverstad, T. R., Gunsolus, J. L., Johnson, G. A. & King, R. P. (2004). Risk-efficiency criteria for evaluating economics of herbicide-based weed management systems in corn. Weed Technology, 18, 687–697. Hrubovcak, J., Vasavada, U. & Aldy, J. E. (1999). Green Technologies for a More Sustainable Agriculture, Agricultural Information Bulletin No. 752. Washington, DC: US Department of Agriculture, Economic Research Service. Hurley, T. M., Mitchell, P. D. & Rice, M. E. (2004). Risk and the value of Bt corn. American Journal of Agricultural Economics, 86, 345–358. Hutchins, S. H. (1997). IPM: opportunities and challenges for the private sector. In Radcliffe’s IPM World Textbook, eds. E. B. Radcliffe & W. D. Hutchison. St. Paul, MN: University of Minnesota. Available at Hutchison, W. D., Burkness, E. C., Carrillo, M. A., Galvan, T. L., Mitchell, P. D. & Hurley, T. M. (2006a). IPM: A Risk Management Framework to Improve Decision-making, Public. No. 08229. St. Paul, MN: University of Minnesota Extension Service. Hutchison, W. D., Burkness, E. C., Carrillo, M. A., Hurley, T. M. & Pahl, G. (2006b). Fresh-Market Cabbage: Increasing Economic Returns while Reducing Risk, Public. No. 08230. St. Paul, MN: University of Minnesota Extension Service. Isard, S. A., Gage, S. H., Comtois, P. & Russo, J. M. (2005). Principles of the atmospheric pathway for invasive species applied to soybean rust. BioScience, 55, 851– 861. Jepson, P. C. (2007). Ecotoxicology: the ecology of interactions between pesticides and non-target organisms. In Perspectives in Ecological Theory and Integrated Pest Management, eds. M. Kogan & P. Jepson, pp. 522–551. Cambridge, UK: Cambridge University Press. Kogan, M. (1998). Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology, 43, 243–270. Kovach, J., Petzoldt, C., Degni, J. & Tette, J. (1992). A Method to Measure the Environmental Impact of Pesticides, New York Food and Life Science Bulletin No. 139. Geneva, NY: Cornell University, New York Agricultural Experiment Station. Available at 5203. Manfredo, M. R. & Leuthold, R. M. (1999). Value-at-risk analysis: a review and the potential for agricultural applications. Review of Agricultural Economics, 21, 99– 111. McCarl, B. & Bessler, D. (1989). Estimating an upper bound on the Pratt risk aversion coefficient when the

utility function is unknown. Australian Journal of Agricultural Economics, 33, 56–63. Mitchell, P. D. (2005). The Expected Net Benefit and Break-Even Probability for Bt Corn in Wisconsin, UWEX Information Bulletin with companion spreadsheet, October. Madison, WI: University of Wisconsin Extension. Available at extension.htm. Mitchell, P. D. (2008). Risk, Farmer Returns and Integrated Pest Management (IPM), Agricultural and Applied Economics Staff Paper No. 526. Madison, WI: University of Wisconsin–Madison. Mitchell, P. D. & Onstad, D. W. (2005). Effect of extended diapause on the evolution of resistance to transgenic Bacillus thuringiensis corn by northern corn rootworm (Coleoptera: Chrysomelidae). Journal of Economic Entomology, 98, 2220–2234. Mitchell, P. D. & Onstad, D. W. (2008). Valuing insect resistance in an uncertain future. In Insect Resistance Management: Biology, Economics, and Prediction, ed. D. W. Onstad, pp. 17–38. San Diego, CA: Academic Press. Mitchell, P. D., Gray, M. E. & Steffey, K. L. (2004). A composed error model for estimating pest-damage functions and the impact of the western corn rootworm soybean variant in Illinois. American Journal of Agricultural Economics, 86, 332–344. Moffitt, L. J., Tanagosh, L. K. & Baritelle, J. L. (1983). Incorporating risk in comparisons of alternative pest management methods. Environmental Entomology, 12, 1003– 1111. Mullen, J. D., Norton, G. W. & Reaves, D. W. (1997). Economic analysis of environmental benefits of integrated pest management. Journal of Agriculture and Applied Economics, 29, 243–253. Mumford, J. D. & Norton, G. A. (1984). Economics of decision making in pest management. Annual Review of Entomology, 29, 157–174. Musser, W. N., Tew, B. V. & Epperson, J. E. (1981). An economic examination of an integrated pest management production system with a contrast between E-V and stochastic dominance analysis. Southern Journal of Agricultural Economics, 13, 119–124. Norton, G. A. (1976). Analysis of decision making in crop protection. Agroecosystems, 3, 27–44. Norton, G. A. (1982). A decision analysis approach to integrated pest control. Crop Protection, 1, 147– 164. Norton, G. A., Heinrichs, E. A., Luther, G. C. & Irwin, M. E. (eds.) (2005). Globalizing Integrated Pest Management: A Participatory Research Approach, Ames, IA: Blackwell.




Nowak, P. (1992). Why farmers adopt production technology. Journal of Soil and Water Conservation, 47, 14–16. Olson, K. D., (2004). Farm Management: Principles and Strategies. Ames, IA: Iowa State University Press. Onstad, D. W. (1987). Calculation of economic injury levels and economic thresholds for pest management. Journal of Economic Entomology, 80, 297–303. Onstad, D. W. & Rabbinge, R. (1985). Dynamic programming and the computation of economic injury levels for crop disease control. Agricultural Systems, 18, 207– 226. Onstad, D. W., Crowder, D. W., Mitchell, P. D. et al. (2003). Economics versus alleles: balancing IPM and IRM for rotation-resistant western corn rootworm (Coleoptera: Chrysomelidae). Journal of Economic Entomology, 96, 1872– 1885. Pedigo, L. P. & Rice, M. E. (2006). Entomology and Pest Management, 5th edn. Upper Saddle River, NJ: Prentice Hall. Pedigo, L. P., Hutchins, S. H. & Higley, L. G. (1986). Economic injury levels in theory and practice. Annual Review of Entomology, 31, 341–368. Peterson, R. K. D. & Hunt, T. E. (2003). The probabilistic economic injury level: incorporating uncertainty into pest management decision-making. Journal of Economic Entomology, 96, 536–542. Plant, R. E. (1986). Uncertainty and the economic threshold. Journal of Economic Entomology, 79, 1–6. Plant, R. E. & Stone, N. D. (1991). Knowledge-Based Systems in Agriculture. New York: McGraw-Hill.

Ragsdale, D. W., McCornack, B. P., Venette, R. C. et al. (2007). Economic threshold for soybean aphid. Journal of Economic Entomology, 100, 1258–1267. Rogers, E. M. (1995). Diffusion of Innovations, 4th edition. New York: The Free Press. Sharpe, W. F. (1994). The Sharpe ratio. Journal of Portfolio Management, 21, 49–58. Shiferaw, B., Freeman, H. A. & Swinton, S. M. (eds.) (2004). Natural Resource Management in Agriculture: Methods for Assessing Economic and Environmental Impacts. Wallingford: UK: CABI Publishing. Shoemaker, C. (1982). Optimal integrated control of univoltine pest populations with age structure. Operations Research, 30, 40–61. Stone, N. D., Gutierrez, A. P., Getz, W. M. & Norgaard, R. (1986). III. Strategies for pink bollworm control in southwestern desert cotton: an economic simulation study. Hilgardia, 54, 42–56. Swinton, S. M. & Day, E. (2003). Economics in the design, assessment, adoption, and policy analysis of IPM. In Integrated Pest Management: Current and Future Strategies, R-140, ed. K. R. Barker. Ames, IA: Council for Agricultural Science and Technology. Venette, R. C. & Ragsdale, D. W. (2004). Assessing the invasion by soybean aphid (Homoptera: Aphididae): where will it end? Annals of the Entomological Society of America, 97, 219–226. Wearing, C. H. (1988). Evaluating the IPM implementation process. Annual Review of Entomology, 33, 17–38.

Chapter 5

IPM as applied ecology: the biological precepts David J. Horn Any insect (or other) pest exists within an ecosystem, consisting of the surrounding biological and physical environment with which it interacts. The interactions between a pest population and its ecosystem are highly complex, and in many cases several pests with different biologies need to be simultaneously managed on a single crop. Ecological issues are exacerbated as the scale of management increases. On a typical farm in midwestern USA we might find fields producing maize (corn), soybeans, hay and perhaps small grains or canola, plus several species of vegetables in a family garden, several kinds of livestock and poultry, stored feed and seed, landscaping plantings, weeds, wildlife and the farmer and his/her household, any and all of which might harbor populations of one or more pests. The farm ecosystem occurs in a matrix of surrounding systems each with its own communities including pests. Ecological processes within surrounding habitats influence events within adjacent areas. In our efforts to maintain high yields and maximize profits, we often oversimplify and override ecosystem processes and unknowingly disrupt whatever naturally occurring pest population regulation there may be. Kogan (1995) and others have noted that even successful IPM programs may pay little heed to the complexity and unpredictability of ecological processes. Our pest management efforts therefore are often disruptive of ecosystem func-

tions. In order to develop more ecologically based IPM systems we need greater understanding of ecological processes. The present chapter introduces some of these fundamental ecological processes as they impact pest populations. While the past few decades have witnessed general acceptance of considering ecology in developing pest management systems, there remains little agreement as to what ecological paradigms are most applicable to pest management systems (Kogan, 1995; Carson et al., 2004). Ecology draws on ideas and data from across biology. As the information base expands, ecological ideas are revised. Ecologists rarely agree completely on such issues as the importance of equilibrium in population regulation, or whether there is a relationship between species diversity and community stability. Those who design and implement pest management systems may be frustrated by the apparent flux in ecological theory. The challenge of applying ecological theory to pest management is intensified by several differences between “natural” ecosystems (such as abandoned fields or forests) and more closely managed, “artificial” ecosystems (such as agricultural fields, orchards or highly manicured landscapes). Ecological ideas generated from studies of natural ecosystems may not be applicable to many managed ecosystems. Even small and isolated ecosystems are enormously complex and variable. Field

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



experiments are subject to widely varying outputs and results may be difficult to interpret. These and other issues supply the context of an ecological approach to IPM.

5.1 Life systems Because a pest population interacts with the surrounding ecosystem it cannot be considered apart from that ecosystem. Clark et al. (1967) developed the “life system” concept to reinforce this reality, and the life system paradigm remains useful 40 years later. A life system consists of the pest population plus its “effective environment,” the sum of the surrounding ecological factors that impact the population positively or negatively. The effective environment includes food, competitors, predators, pathogens and the physical surroundings – in short, anything that reduces or increases survivorship, fecundity and movement of the pest (or any other organism). The scale of the life system is arbitrary and the impact on the effective environment varies over space and time; the life system will be different according to whether one views the surrounding ecosystem as a single plant, a single field or orchard, a regional landscape or an ecoregion. An applied ecological approach to IPM needs to consider an appropriate scale. Insects such as the alfalfa snout beetle (Otiorhynchus ligustici) that do not fly may be limited to a local area for many decades whereas to understand the ecology of long-distance migrants such as the beet armyworm (Spodoptera exigua) or migratory locust (Schistocerca gregaria) may require consideration of an effective environment over hundreds of square kilometers. In most cases the effective environment needs to be considered at least to the level of the local landscape (Duelli, 1997; Collins & Qualset, 1999; Landis et al., 2000). An implication of the life system concept is that agricultural or forestry activities such as tilling, thinning or harvesting either disrupt or enhance ecosystem functions resulting in a more or less favorable environment for a pest population. This in turn leads to either an increase or a decrease of its population. Such management activities may have significant unintended influence on the most carefully designed ecological IPM systems.

5.2 Pest population dynamics One approach to improve our understanding of the complex interaction between a pest population and its effective environment is to use relatively simple population models. Even highly simplified population models can provide a variety of outputs illustrating general ecological principles. As an introductory example, we can denote population density with a single value N and (initially) ignore the obvious fact that members of a population vary regarding a wide variety of biological traits (see section “Population structure and life tables” below). Here I use the single term N for convenience to denote population density, although it actually represents a range of individuals assumed identical only for study and preliminary analysis. To increase realism, Ehrlich et al., 1975 suggested that all populations have the following general characteristics. (1) Populations and their effective environments are always changing in time and space, and a model of a population at a single location and time interval likely does not adequately represent events even in the same population at another place and time. The concept of the “metapopulation” (discussed below) addresses this issue. (2) Practical realities may dictate that we need to manage localized populations (e.g. within a single infested field) but management plans should be developed after an understanding of the pest’s population dynamics on a landscape level or (ideally) over its entire geographic distribution. (3) Variation in demographic characteristics (birth rate, emigration, etc.) within a local population may exceed variation in the same parameters among adjacent or distant populations of the same species. (4) Immigration does not always increase gene flow and changes in gene frequency do not necessarily follow immigration. For instance, corn earworm (Helicoverpa zea) moving from south to north in the USA may or may not carry genes for insecticide resistance resulting from intensive insecticide use on crops in which they originated. The Lotka–Volterra model of logistic population growth (Lotka, 1920) remains a useful mathematical model to illustrate the role of equilibrium in population dynamics. This classical model










400.0 200.0

800.0 600.0 400.0 200.0






30 5

Time (generations)






1600.0 1400.0 Abundance

1200.0 1000.0 800.0 600.0 400.0 200.0 5


15 20 Time (generations)





2000.0 Abundance

presupposes that many populations tend to be regulated about an equilibrium set by their effective environment (the causes for equilibrium may not be clear in many cases, and the model may be misleading in a practical sense, as many populations only appear to be in equilibrium). In the Lotka– Volterra model, K represents the environmental carrying capacity and acts to regulate population growth according to the following relationship:

where N = population density

1500.0 1000.0 500.0

t = time interval b = birth rate d = death rate K = carrying capacity.



15 20 Time (generations)




where r = (b – d) and e = base of natural logarithms.) The difference and differential forms of the model describe essentially the same phenomenon, but the difference equation form is capable of a large array of outputs due to impact of time delays (Horn, 1988). Figure 5.2 illustrates this. From an IPM perspective, if b is large relative to d (characterizing a population with a high intrinsic rate of increase), there is a tendency for great oscillations about K, resulting in apparent



2000.0 Abundance

This equation gives the familiar and intuitively satisfying sigmoid curve of population growth (Fig. 5.1). (The above is a difference equation; in discussion of ecological models often seen in textbooks, the model may be in differential form, integrated to: N = K/(1 − e−r t )


Time (generations)

Fig. 5.1 Logistic population growth for a population with r = 1.5, K = 1000 and initial population = 100. Simulation using software from Akc¸akaya et al. (1999).

N t+1 − N t = N t (b − d)(K − N t )/K


1500.0 1000.0 500.0



15 20 Time (generations)



Fig. 5.2 Results of simulations for populations with K = 1000 and initial population = 100 (as in Fig. 5.1) and progressively increasing growth rate. (A) r = 6; (B) r = 10; (C) r = 15; (D) r = 18. Simulation using software from Akc¸akaya et al. (1999).

instability even though there is equilibrium in the mathematical sense. A population with high r may approach K with such speed that at the next iteration it exceeds K and the population then declines, precipitously if N > K




(Fig. 5.2B). As one increases b while holding d and K constant in computer simulations of this model, one obtains an array of results: low-amplitude cyclic oscillations about K, high-amplitude stable limit cycles (Fig. 5.2C) and/or cycles whose periodicity cannot be distinguished from random (Fig. 5.2D). From a modeling standpoint, these results are simple functions of the ratio of b to d and/or the relationship of the initial N to K. Such outputs do reflect actual results of studies on local populations of arthropods with high fecundity and short generation time, such as aphids, spider mites and houseflies. The model thus predicts that local insect populations with high fecundity and short generation time may fluctuate wildly and unpredictably, never appearing to be in equilibrium locally but rather exhibiting spectacular instability. Such populations also reach economically damaging levels much more rapidly than do those with lower r. The array of adaptations (high fecundity, short generation time, low competitive ability and high dispersal) has been termed “r-selection” (MacArthur & Wilson, 1967), and results from selection in environments favoring maximum growth, such as temporarily available habitats that occur early in ecological succession. On a local level such r-selected species may quickly over exploit their resources and local extinction follows. The two-spotted spider mite (Tetranychus urticae) is an example; females lay a large number of eggs and the resulting nymphs can reach adulthood in six days. Spider mites can overwhelm and destroy an untended house plant in a few weeks. In contrast to r-selection, “K-selection” is typical of species that occupy habitats with longer spatial and temporal stability. These species exhibit lower fecundity, longer generation times, low dispersal tendency and high competitive ability. Although not an agricultural pest, the tsetse flies (Glossina spp.) are examples of extreme Kselection; females give birth to one mature larva at a time after a long incubation period. Characteristics of both r- and K-selection may occur within a species and may show seasonal variation. Saltmarsh planthoppers (Prokelisia spp.) may be short-winged during the growing season. Dispersal is limited and high fecundity results from feeding on vigorous hosts. As their food supply

dwindles later in the season, long-winged morphs develop forms with lower fecundity and highly dispersive, relatively K-selected (Denno, 1994). Typically, agricultural pests exhibit adaptations consistent with r-selection. Annual agricultural crops are disrupted due to tilling and harvesting and ecological succession is frequently reset to its starting point annually. This seems to select in favor of phytophagous insects that can locate and exploit their food quickly and efficiently. Adaptations of colonizing species of plants and insects are consistent with r-selection, i.e. efficient dispersal and a tendency to increase numbers quickly after locating favorable habitat. Many crop plants (or their ancestors) are typical of the pioneering stages of plant succession, as are their associated insect pests. Conventional agriculture, especially when undergoing frequent tillage, invites early-successional pest species that are more likely to exhibit outbreaks simply due to their r-selected adaptations. Populations of such pests may not display equilibrium over the time interval that the crop is in production. There may not be enough time during the growing season for the population to increase to its carrying capacity. The model describes this situation with high r, i.e. birth rate greatly exceeds death rate (until harvest, when the insects all emigrate or die). Equilibrium around K might be more likely in longerlasting systems such as forests and orchards. Also, pest population fluctuations in these more complex ecosystems are partly buffered by the complexity of interactions within food webs, and outbreak of any particular pest species is less likely (see below under “Population stability and species diversity”). The logistic model describes density-dependent population regulation, which by definition is the major way to regulate a population about equilibrium. The effect of a density-dependent regulating factor is a function of the numbers (N) within a population. At low density the impact is less while at high density the impact is extreme. Densitydependent factors include predation, parasitism and competition. Density-independent factors, by contrast, may control a population but by definition do not regulate. Weather, pesticides, tillage and harvesting are all density-independent factors. In the life systems of most pest populations,


5.3 Stochastic models and metapopulations The more realistic a model is in accurately describing real events, the more complicated it usually is. Stochastic models assign probability functions to birth rate, death rate, carrying capacity and the other components of the life system, so that each term in the logistic model above becomes a probability function that represents fluctuations about a mean (Fig. 5.3). Such probabilistic models are less mathematically tractable and results may be less intuitively understandable than are deterministic models, but such models do supply greater realism in describing actual population events and can be more useful in insect pest management. Use of computer simulations has removed a major obstacle to applying stochastic models in pest management, although experimental validation of such models is tedious (Pearl et al., 1989). To address the shortcomings of local equilibrium models in describing population events over an area of interest beyond a single crop to the regional and landscape levels, we may consider the demography of populations of the same species throughout a region by including disper-

A 1000.0


800.0 600.0 400.0 200.0



20 15 Time (generations)





2000.0 Abundance

both density-dependent and density-independent factors have impacts and the relative importance of each differs, sometimes leading to an impression that one or the other is the dominant or even the exclusive influence in determining population density (Horn, 1968). Density-dependent models assume that equilibrium exists and that one among many factors is the one that regulates (Hunter, 1991). Others (e.g. Strong, 1986; Stiling, 1988) argue that density dependence may not be very important in determining numbers of most animal populations. Chesson (1981) suggested that density-dependent regulation might occur most often at extremes of abundance and that in most populations at medium ranges of density the influence of density-dependent regulating factors was undetectable. Many more recent studies of natural populations seem to support this view (Cappuccino & Price, 1995; Frank van Veen et al., 2006).

1500.0 1000.0 500.0



15 20 Time (generations)



Fig. 5.3 Results of stochastic simulation of logistic growth model, ten iterations. Solid line is mean value; vertical lines represent standard deviation. (A) r = 1.5 (as in Fig. 5.1); (B) r = 18 (as in Fig. 5.2D). Simulation using software from Akc¸akaya et al., (1999).

sal as a factor involved in population regulation. The interacting system of local populations over a region is a metapopulation (Gilpin & Hanski, 1991). The metapopulation occupies both favorable and unfavorable places. Where the environment is favorable (a “source” area), the population is usually increasing (b > d) and the excess disperses to other regions including “sink” areas, where b < d. The local population in a sink would disappear were it not supplemented by immigration. The relationship among sources, sinks and dispersal maintains the population in a more or less steady state. Movement among sources and sinks can create an impression that the metapopulation is in equilibrium throughout a wide area even though there is no local equilibrium evident in any one area (Murdoch, 1994). Usually our local area occupies our greatest interest when we apply IPM practically. The ideal goal of IPM is to reduce pest numbers more or less permanently below the economic injury level (EIL). This implies that equilibrium exists. The classic EIL model (Stern et al., 1959) depends on models of population dynamics that presuppose equilibrium. These models are



Table 5.1

Hypothetical life table for an insect


Mortality Survivorship Mortality rate

Eggs 1000 0 0.00 Larvae 1000 700 0.70 Pupae 300 120 0.40 Reproducing 120 24 0.20 adult females Old adults 24 24 1.00 Average number of eggs produced per female = 50


400.0 300.0 200.0 100.0






Time (generations) B 500.0 400.0 Abundance




300.0 200.0 100.0

understandable, tractable and intuitively satisfying, but in real populations there may not be a general equilibrium position for density of many and perhaps most pest species.






Time (generations) C

500.0 400.0

5.4 Population structure and life tables



300.0 200.0 100.0

As noted above, individuals in a population vary so the use of a single term N to denote a population is an oversimplification that can be misleading. One tool to illustrate the importance of such issues as age structure and reproductive ability is through the use of life tables. A life table is a summary of the vital statistics of a population. Life tables can be highly detailed and very useful analytical tools but to construct one involves a huge amount of sampling. Table 5.1 presents an oversimplified, hypothetical life table for illustration. The life table contains a census of the numbers in each age class (here, instars), along with survivorship, mortality and mortality rate within each age class. Having a life table allows us to conduct simulations to predict the impact of various mortality factors. This is illustrated in Fig. 5.4. Clearly mortality among adults has a greater impact on population persistence in these examples. Obviously a population initially consisting entirely of reproducing females increases at a much higher rate than does a population dominated by non-reproductive individuals. To most quickly bring a population under control,






Time (generations)

Fig. 5.4 Simulations based on hypothetical life tables, initial population of 10 reproducing females. (A) Mortality rates as in Table 5.1; population numbers level at 100; (B) as in Table 5.1 but larval mortality halved and adult mortality doubled; population extirpated after 5 generations; (C) as in Table 5.1 but larval mortality doubled and adult mortality eliminated; population numbers level at 500. Simulation using software from Akc¸akaya et al. (1999).

pre-reproductive females should be targeted to increase mortality.

5.5 Population stability and species diversity As noted, the effective environment includes competitors, predators, parasites, pathogens – all ecosystem components that have an impact on a particular species. Food webs for even very simple habitats exhibit a large array of species. The


invertebrate fauna associated with cabbage in Minnesota includes at least 11 leaf feeders, 10 sap feeders, 4 root feeders, 21 feeders on decaying plant matter and 79 saccharophiles (feeding on sugar produced by the plant or by sapfeeding Hemiptera) for a total of 125 species of primary consumers (herbivores in the most inclusive sense) (Weires & Chiang, 1973). Additionally Weires and Chiang documented 85 species of arthropod predators and parasitoids. This local diversity is not unusual; I have recorded well over 1200 species of arthropods in my modest urban house and yard. At the crop ecosystem or landscape level the diversity is bewildering and again we may resort to simplified models to improve understanding. Species diversity is a function of the numbers and proportion of each species present in an ecosystem. It is assumed that species diversity in turn reflects the actual number of interconnections in a food web. Biodiversity has become a popular term in discussions about complexity in agricultural and other human-dominated ecosystems (Stinner et al., 1997). Altieri & Nicholls (1999) defined biodiversity as “all species of plants, animals and microorganisms existing and interacting within an ecosystem.” In agroecosystems this includes phytophagous species (pests and non-pests), predators, parasitoids, parasites, pollinators and decomposers. Debate continues among theoretical and applied ecologists as to the nature of any relationship between species diversity and stability of individual populations within an ecosystem. It is often assumed that pest outbreaks are generally suppressed in more complex (and therefore more diverse) ecosystems. This so-called “diversity–stability hypothesis” holds that communities with higher species diversity (greater biodiversity) are more stable because outbreaks of any particular species are ameliorated by the checks and balances and alternative food web pathways that exist within a large and integrated ecosystem. Andow (1991) assembled an exhaustive list of studies of agricultural systems addressing the diversity–stability hypothesis and he found that herbivores were less abundant in diverse plantings in 52% of cases. Most of these studies intermixed other plant species with the primary host of a specialist herbivore,

leading to reduced populations of the specialist herbivore (Risch et al., 1983; Altieri, 1994). Root (1973) formulated this as the resource concentration hypothesis: herbivores are more likely to find and remain on hosts that are growing in dense or nearly pure stands; the most specialized species frequently attain higher densities in simple environments. As a result, biomass tends to become concentrated in a few species, causing a decrease in the diversity of herbivores in pure stands.

Observed increases in herbivore populations in crop monocultures seem to result from higher rates of colonization and reproduction along with reduced emigration, predation and parasitism. Other studies (e.g. Tilman et al., 1996) have shown experimentally that soil nutrients are more completely cycled and productivity increases in more diverse ecosystems. Structural diversity (Andow & Prokrym, 1990) is important too; cropping systems with taller plants (such as maize among cucurbits and beans) provide more physical space to arthropods and this seems to increase diversity at all trophic levels (Altieri, 1994). A positive relationship between diversity and stability is intuitively satisfying yet difficult to prove experimentally. Altieri & Nicholls (1999) believed that as biodiversity increased within an agroecosystem, more internal links develop within food webs and these links result in greater population stability and fewer pest outbreaks. This presupposes that most of the interconnecting trophic web is comprised of density-dependent links, which may or may not be the case. The detailed trophic structure of agricultural systems has only recently begun to be analyzed at sufficient detail to adequately address this question (e.g. Janssen & Sabelis, 2004; Jackson et al., 2007). Southwood & Way (1970) suggested that stability of insect populations in agroecosystems depended on the “precision” of density-dependent responses within the food web, and the precision in turn depends on four major ecosystem parameters: (1) surrounding and within-crop vegetation, (2) permanence of the crop in space and time, (3) intensity of management including frequency of disruptions like pesticide applications and tillage and (4) degree of isolation of the agroecosystem from surrounding unmanaged vegetation.




The overall stability of any ecosystem is certainly a function of the sum of interactions among plant, pests, natural enemies and pathogens, but the specific details are often elusive. Vandermeer (1995) proposed that biodiversity in agroecosystems has two components, planned and associated. Planned biodiversity consists of cultivated crops, livestock and other organisms (such as biological control agents) that are purposely introduced into an agroecosystem for direct economic benefit. Planned biodiversity is normally intensively managed, primarily to produce high yields. Associated biodiversity includes all the other plants, herbivores, carnivores and microbes that either pre-exist in or immigrate into the agroecosystem from the surrounding landscapes. Whether associated biodiversity persists within an agroecosystem depends on whether the ecological requirements of each organism are met over the time that the agroecosystem exists. Vandermeer (1995) proposed that a high amount of associated biodiversity is essential to maintaining stability of arthropod populations that otherwise might negatively impact planned biodiversity. Consideration of the relationship between planned and associated biodiversity is helpful in developing pest management practices that enhance overall biodiversity. This can lead to greater sustainability due to increased impact of biological control agents, reduced soil loss and enhanced on-site nutrient cycling.

5.6 Open and closed ecosystems Open (or subsidized) ecosystems depend on importation of nutrients and energy from outside, whereas closed ecosystems are relatively selfcontained with respect to nutrients and energy. From open ecosystems there is periodic removal of a large proportion of nutrients. A field of maize under conventional tillage is an example, with heavy importation of fertilizer at planting and subsequent energy inputs associated with tilling, pesticide application and so forth (over 20 million tonnes of chemical fertilizer are used annually in the USA, much of it on maize). Yield and crop residue comprise most of the nutrients in a maize field and these are removed at harvest. In

addition, the species comprising the food web in maize are largely artificial and recent; interspecific associations are not long-standing and little coevolution has occurred. Maize is native to Mesoamerica whereas many of its major insect pests, e.g. European corn borer (Ostrinia nubilalis) are from other regions of the world. It takes a long time for native natural enemies to expand their host or prey range to include exotic organisms. In a closed ecosystem, such as a deciduous forest, much nutrient cycling occurs on-site with far less importation from elsewhere. The nutrients and energy in the canopy and understory fall to the ground as leaves, insect frass or dead insects, or are converted into arthropods which are eaten by larger predators. Much of the flora and fauna does not leave the ecosystem. Migratory species do leave, of course, but often they return. Overall, there is rather little movement of nutrients into and out of the system. Many closed ecosystems (like the deciduous forest) have existed as assemblages of species for millennia, and the resulting food webs contain many coevolved trophic links and close ecological associations among mostly native species. Such ecosystems can be somewhat resistant to invasion by exotic species. Of course, there are exceptions, as the gypsy moth (Lymantria dispar) and emerald ash borer (Agrilus planipennis) have demonstrated in forests of the USA. Most agroecosystems are open ecosystems and artificial assemblages, and because our main objective is to extract a usable product at a profit, it is often economical to maintain these systems in their present state despite a potential increase in pest activity (Lowrance et al., 1984). Even in an artificial landscaped ecosystem that we do not harvest, we fertilize (providing nutrient input) and remove fallen leaves and lawn waste, exporting productivity in order to maintain a pleasing appearance. The species assemblage in planned landscapes often includes a preponderance of exotic species; for instance, from my home in Ohio I can view Chinese ginkgo, Colorado blue spruce, English walnut, Norway maple and Siberian elm trees. “Open” and “closed” are arbitrary designations and the two types of ecosystems are points on a continuum. There are distinct differences between the two in the level of impact of pests


and of management procedures (Altieri, 1987, 1994). Especially in commercial agriculture, open and simplified agroecosystems are often devoted to a single crop resulting in higher pest populations and reduced species diversity. Generally the greater the modification and perturbation in the direction of ecosystem simplification and energy subsidy, the more abundant are insect pests. These reductions in biodiversity may extend beyond the local level, impacting the normal functioning of surrounding ecosystems with potential negative consequences for successful IPM (Flint & Roberts, 1988).

5.7 Scale and ecologically based pest management As mentioned earlier, the scale of the area involved is an important consideration in assessing a pest problem and developing an ecological approach to its management. The perception of pest problems, estimation of economic injury levels and approach to pest management can vary greatly depending on scale. One can view an ecosystem at the level of an individual plant, a small research plot, a large crop field or an entire farm with its regional “agropastoral” (Altieri, 1994) landscape, the local watershed and so forth. The wider the area under consideration, the more complex the interactions are likely to be (Wilkinson & Landis, 2005; Marshall et al., 2006). At the metapopulation level the regional dynamics of a life system may be very different from the local dynamics, and as noted earlier a pest population may appear to be in equilibrium throughout a regional landscape, even when there is no discernible equilibrium at the local level of interest to a pest manager. Results from small plot research may not be applicable to a larger scale (Kemp et al., 1990). Local movement of pests may be less important at smaller scales (e.g. the individual plant) but very influential in population dynamics at a regional landscape level especially if this includes surrounding unmanaged habitat. For instance, Mexican bean beetle (Epilachna varivestis) overwinters in hedgerows and along field edges, so that soybean fields closest to these overwintering

sites are likely to become infested earlier and subsequent Mexican bean beetle populations will be higher. Soybeans located near pole and bush beans (more favorable hosts for the Mexican bean beetle) are more likely to develop economically important infestations earlier (Stinner et al., 1983). Natural enemies often move from unmanaged edge habitat and nearby abandoned fields and forests into adjacent farm fields and the nature of this movement may be very important to local suppression of pests (Wilkinson & Landis, 2005). An increase in intensive agriculture often leads to reduction in the size and diversity of surrounding unmanaged communities with their rich store of associated biodiversity including natural enemies. Elimination of hedgerows at field edges in the USA is an example. Many studies have shown that there are increased numbers and activity of predators and parasitoids near field borders when there is sufficient natural habitat to provide cover and alternate prey and hosts, as well as nectar and pollen for adult parasitoids to eat. Unmanaged border areas significantly enhance biological control (van Emden, 1965; Marino & Landis, 1996).

5.8 Conclusions All ecosystems are complex, even apparently “simplified” monocultures. While simple intuitive equilibrial models may help us to understand ecological processes, the actual impact of alternate crops, weeds, competitors, natural enemies and other associated organisms on the life systems of pest species may be inconsistent and highly variable. The challenge is to develop ecological approaches to pest management in the face of this complexity, and to proceed with an appreciation of ecological processes despite gaps in our understanding.

References Akc¸akaya, H. R., Burgman, M. A. & Ginzburg, L. R. (1999) Applied Population Ecology Principles and Computer Exercises. Sunderland, MA: Sinauer Associates. Altieri, M. A. (1987). Agroecology: The Scientific Basis of Alternative Agriculture. Boulder, CO: Westview Press.




Altieri, M. A. (1994). Biodiversity and Pest Management in Agroecosystems. New York: Food Products Press. Altieri, M. A. & Nicholls, C. I. (1999). Biodiversity, ecosystem function, and insect pest management in agricultural systems. In Biodiversity in Agroecosystems, eds. W. W. Collins & C. O. Qualset, pp. 69–84. Boca Raton, FL: CRC Press. Andow, D. A. (1991). Vegetational diversity and arthropod population response. Annual Review of Entomology, 36, 561–586. Andow, D. A. & Prokrym, D. R. (1990). Plant structural complexity and host finding by a parasitoid. Oecologia, 62, 162–165. Cappuccino, N. & Price, P. W. (eds.) (1995). Population Dynamics: New Approaches and Synthesis. New York: Academic Press. Carson, W. P., Cronin, J. P. & Long, Z. T. (2004) A general rule for predicting when insects will have strong top down effects on plant communities: on the relation between insect outbreaks and host concentration. In Insects and Ecosystem Function, eds. W. W. Weisser & E. J. Siemann, pp. 193–211. Berlin, Germany: SpringerVerlag. Chesson, P. L. (1981). Models for spatially distributed populations: the effect of within-patch variability. Theoretical Population Biology, 19, 288–325. Clark, L. R., Hughes, R. D., Geier, P. W. & Morris, R. F. (1967). The Ecology of Insect Populations in Theory and Practice. London: Methuen. Collins, W. W. & Qualset, C. O. (eds.) (1999). Biodiversity in Agroecosystems. Boca Raton, FL: CRC Press. Denno, R. F. (1994). Life history variation in planthoppers. In Planthoppers, Their Ecology and Management, eds. R. F. Denno & T. J. Perfect, pp. 163–215. New York: Chapman & Hall. Duelli, P. (1997). Biodiversity evaluation in agricultural landscapes: an approach at two different scales. Agriculture Ecosystems and Environment, 62, 81–91. Ehrlich, P. R., White, R. R., Singer, M. C., McKechnie, S. W. & Gilbert, L. I. (1975). Checkerspot butterflies: a historical perspective. Science, 188, 221– 228. Flint, M. L. & Roberts, P. A. (1988). Using crop diversity to manage pest problems: some California examples. American Journal of Alternative Agriculture, 3, 164– 167. Frank van Veen, F. J., Morris, R. J. & Godfray, H. C. J. (2006). Apparent competition, quantitative food web shifts and the structure of phytophagous insect communities. Annual Review of Entomology, 51, 187– 208.

Gilpin, M. & Hanski, I. (1991). Metapopulation Dynamics: Empirical and Theoretical Investigations. San Diego, CA: Academic Press. Horn, H. S. (1968). Regulation of animal numbers: a model counter-example. Ecology, 49, 776–778. Horn, D. J. (1988). Ecological Approach to Pest Management. New York: Guilford Press. Hunter, A. F. (1991). Traits that distinguish outbreaking and non-outbreaking Macrolepidoptera feeding on northern hardwood trees. Oikos, 60, 275– 282. Jackson, L. E., Pascual, U. & Hodgkin, T. (2007), Utilizing and conserving agrobiodiversity in agricultural landscapes. Agriculture Ecosystems and Environment, 121, 196–210. Janssen, A. & Sabelis, M. W. (2004). Food web interactions and ecosystem processes. In Insects and Ecosystem Function, eds. W. W. Weisser & E. J. Siemann, pp. 193–211. Berlin, Germany: Springer-Verlag. Kemp, W. P., Harvey, S. J. & O’Neill, K. M. (1990). Patterns of vegetation and grasshopper community composition. Oecologia, 83, 299–308. Kogan, M. (1995). IPM: historical perspectives and contemporary development. Annual Review of Entomology, 43, 243–270. Landis, D. A., Wratten, D. & Gurr, G. M. (2000). Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175–201. Lotka, A. J. (1920). Analytical notes on certain rhythmic relations in organic systems. Proceedings of the National Academy of Sciences of the USA, 7, 410–415. Lowrance, R., Stinner, B. R. & House, G. J. (eds.) (1984). Agricultural Ecosystems: Unifying Concepts. New York: John Wiley. MacArthur, R. H. & Wilson, E. O. (1967). The Theory of Island Biogeography. Princeton, NJ: Princeton University Press. Marino, P. C. & Landis, D. A. (1996). Effects of landscape structure on parasitoid diversity in agroecosystems. Ecological Applications, 6, 276–284. Marshall, E. J. P., West, T. M. & Kleijn, D. (2006). Impacts of an agri-environment field margin prescription on the flora and fauna of arable farmland in different landscapes. Agriculture Ecosystems and Environment, 113, 36–44. Murdoch, W. W. (1994). Population regulation in theory and practice. Ecology, 75, 271–287. Pearl, D. K., Bartoszynski, R. & Horn, D. J. (1989). A stochastic model for simulation of interactions between phytophagous spider mites and their


phytoseiid predators. Experimental and Applied Acarology, 7, 143–151. Risch, S. J., Andow, D. & Altieri, M. A. (1983). Agroecosystem diversity and pest control: data, tentative conclusions and new research directions. Environmental Entomology, 12, 625–629. Root, R. B. (1973). Organization of a plant–arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecological Monographs, 43, 95–124. Southwood, T. R. E. & Way, M. J. (1970). Ecological background to pest management. In Concepts of Pest Management, eds. R. L. Rabb & F. E. Guthrie, pp. 6–29. Raleigh, NC: North Carolina State University Press. Stern, V. M., Smith, R. F., van den Bosch R. & Hagen, K. S. (1959). The integration of chemical and biological control of the spotted alfalfa aphid: the integrated control concept. Hilgardia, 29, 81–101. Stiling, P. (1988). Density-dependent processes and key factors in insect populations. Journal of Animal Ecology, 57, 581–594. Stinner, R. E., Bayfield, C. S., Stomach, J. L. & Dose, L. (1983). Dispersal and movement of insect pests. Annual Review of Entomology, 28, 319–335. Stinner, D. H., Stinner, B. R. & Martsolf, E. (1997). Biodiversity as an organizing principle in agroecosystem

management: case studies of holistic resource management practitioners in the USA. Agriculture Ecosystems and Environment, 62, 199–213. Strong, D. R. (1986) Density-vague population change. Trends in Ecology and Evolution, 1, 39–42. Tilman, D., Walden, D. & Knops, J. (1996). Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature, 379, 718–720. Vandermeer, J. (1995). The ecological basis of alternative agriculture. Annual Review of Ecology and Systematics, 26, 201–224. van Emden, H. F. (1965). The role of uncultivated land in the biology of crop pests and beneficial insects. Scientific Horticulture, 17, 121–126. Weires, R. W. & Chiang, H. C. (1973). Integrated Control Prospects of Major Cabbage Insect Pests in Minnesota: Based on the Faunistic, Host Varietal and Trophic Relationships, Technical Bulletin No. 291. St. Paul, MN: University of Minnesota, Agricultural Experiment Station. Wilkinson, T. K. & Landis, D. A. (2005). The role of plant resources and habitat diversification in biological control. In Plant Provided Food and Plant–Carnivore Mutualism, eds. F. L. Wackers, P. C. J. Van Rijn & J. Bruin, pp. 305–325. Cambridge, UK: Cambridge University Press.


Chapter 6

Population dynamics and species interactions William E. Snyder and Anthony R. Ives

Agricultural monocultures are often thought to be more prone to herbivore outbreaks than natural systems, and early agroecologists posited that the lack of biodiversity in agricultural systems contributes to their instability (Pimentel, 1961; van Emden & Williams, 1974). In contrast, some detailed reviews have concluded that perhaps one or two particularly effective natural enemies are all that is needed for effective pest control (Hawkins et al., 1999). Such issues come to the fore when a decision must be made in classical biological control about whether to introduce one or several natural enemy species in an effort to control exotic pests (Myers et al., 1989; Denoth et al., 2002), and when designing schemes to conserve indigenous natural enemies by modifying cultural practices (Landis et al., 2000; Tscharntke et al., 2005). Here, we first review the major classes of natural enemies – specialists and generalists – and the traits of each that are likely to contribute to (or detract from) their effectiveness as biological control agents. We then discuss interactions within diverse communities of natural enemies that are likely to affect biological control.

6.1 Specialist natural enemies: the best biological control agents? Biological control practitioners have long debated the question: what are the traits of an effective biological control agent? General consensus seems to focus around a few traits that a successful agent will possess (see Chapter 9). For example, Debach & Rosen (1991) describe three key attributes of an effective natural enemy that we condense to: (1) a high degree of prey specificity; (2) a reproductive rate as high as that of the target prey (the pest); and (3) a tolerance of environmental conditions similar to that of the pest. The reason for the third of these attributes is obvious, as a biological control agent must be able to survive in the same environment where the target pest occurs. However, the logic underlying the first two attributes is more complex and subtle.

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


The first attribute describes a biological control agent that is a specialist attacking only one (or perhaps just a few) prey species. The second attribute describes a natural enemy that has a reproductive rate as rapid as its prey, so that the natural enemy is able to reproduce at a pace similar to that of the pest. These traits together can allow a natural enemy to exert density-dependent population regulation, in which mortality from the specialist increases as pest populations grow through time, thereby squelching pest outbreaks (Hassell, 1978; Turchin, 2003). Below, we give a more detailed discussion of the disadvantages and advantages of specialist natural enemies. We focus on parasitoids, which in the literature on the biological control of insect pests are almost synonymous with specialist natural enemies. We will describe the disadvantages and advantages of specialist biological control agents using two well-studied systems to carry the discussion: pea aphid (Acrythosiphon pisum) in alfalfa and red scale (Aonidiella aurantii) in citrus crops.

6.1.1 Boom-and-bust predator–prey cycles Although many specialist natural enemies show density-dependent responses to increasing pest populations, this generally occurs with a time lag roughly equal to the generation time of the specialist parasitoids. This time lag occurs because it takes a generation for increasing pest abundance to give rise to new adult parasitoids. The time lag can cause boom-and-bust cycles in pest densities (Nicholson & Bailey, 1935; Huffaker, 1958; Hassell, 1978); as pest populations grow, they can reach high densities before the parasitoid population becomes large enough to suppress the pest, but once suppression starts, it continues until the pest population is reduced to very low abundances. The tendency of specialist parasitoids to cause boom-and-bust cycles with their prey is illustrated by a simple experiment with pea aphid feeding on alfalfa (Ives, 1995). Pea aphids are attacked by the parasitoid Aphidius ervi, a solitary braconid wasp that was introduced into North America as a biological control agent of pea aphid (Mackauer & Kambhampati, 1986). In greenhouse cages, we introduced pea aphids and then A. ervi adults into each of two cages containing 48 alfalfa plants

Fig. 6.1 Pea aphid density (solid lines) and parasitism by A. ervi (dashed lines) in greenhouse cages (A) without and (B) with bridges connecting 48 alfalfa plants. Pea aphid density is the average number of aphids per plant times 0.025 to fit within the same panels as proportion parasitism. Proportion parasitism is measured by the ratio m/(m + a) where m is the number of new mummies observed between samples taken 2 days apart, and a is the number of pea aphids. Data from Ives (1995).

(Fig. 6.1). The cages differed, with one cage having wire “bridges” between plants on which nonwinged aphids could cross, while the other cage had barriers between plants that limited betweenplant aphid movement to winged aphid adults. The bridges allowed aphids to disperse more evenly among plants, which increased the rate of aphid population growth. In both cages, pea aphid populations continued to increase for about 15 days following parasitoid introduction, which is roughly one parasitoid generation. After the aphid population peaked it declined rapidly, with the aphid population crashing as parasitism reached 100%. Although bridges among plants increased the population growth rate of the aphids, this did not prolong the time until their population crashed; in fact, it sped the rate of collapse. Aphidius ervi was clearly a very effective natural enemy, able to kill all pea aphids in the cages




very quickly regardless of the initial growth rate of the aphid population. This control was density dependent, because percent parasitism increased as pea aphid density increased through time. The success of A. ervi can be attributed to the first two attributes for an effective biological control agent listed above: high specificity and rapid population growth. However, although A. ervi did exterminate pea aphids from the cages, this did not occur before aphids reached high densities, increasing for 12–20 days following the introduction of parasitoids. Furthermore, the extinction of pea aphids from cages naturally caused the extinction of parasitoids. In the field, this would then make the crops susceptible to subsequent pest invasions. Thus, high levels of parasitism came at the expense of stability, with the specialist parasitoid generating boom-and-bust population cycles (Hassell & May, 1973).

6.1.2 Stable pest control by specialists Theoretical models have identified several mechanisms that allow highly specialized parasitoids with high reproductive rates to nonetheless provide stable control of host (pest) populations, without boom-and-bust cycling (Briggs & Hoopes, 2004). A well-studied mechanism is parasitoid spatial aggregation (Hassell & May, 1974; Hassell et al., 1991; Ives, 1992b). If pests are distributed in a patchy fashion, for example in colonies on individual plants, then parasitoids must search among patches for pests. If parasitoid searching causes some hosts to be more or less susceptible to parasitism, then host–parasitoid dynamics can be stabilized. This occurs when, for example, parasitoids aggregate in patches with high host density (Hassell & May, 1974). It also occurs when parasitoids search independently of host density but nonetheless aggregate their searching in a subset of patches (Hassell et al., 1991; Ives, 1992b). The reason this mechanism stops host–parasitoid cycles is that, by focusing parasitoid attacks on a subset of hosts, it increases competition among parasitoids as percent parasitism increases. This stops parasitoid populations from reaching very high densities that then lead to host population crashes (Hassell et al., 1991). There is a cost to this stabilizing effect of parasitoid aggregation, however; due to parasitoid competition, the average host pop-

ulation size is relatively high when there is parasitoid aggregation, creating a trade-off between the effectiveness of host suppression and the ability of the parasitoid to maintain stable host population densities (Ives, 1992b). Despite a lot of interest in the ecological literature, we know of no clear empirical demonstrations of aggregation stabilizing host–parasitoid dynamics, and there are considerable logistical difficulties to testing this idea (Gross & Ives, 1999; Olson et al., 2000). Another mechanism enabling parasitoids to achieve low and stable pest population control without cycling is selective parasitism on only some of the developmental stages of hosts, leaving a long period of invulnerability during development (Murdoch et al., 1992, 2003). Murdoch and colleagues (2005) have demonstrated the ability of the parasitoid Aphytis melinus to control California red scale experimentally in the field (Fig. 6.2A). They caged in entirety 14 citrus trees, and in four cages they experimentally augmented the abundance of red scale to 80 times the abundance in control cages. Within two red scale generations, experimentally augmented populations were brought back to control cage levels as parasitism rates reached 95% (compared to 66% in controls). Using a detailed, system-specific model that includes all developmental stages of red scale, Murdoch and colleagues (2005) demonstrated two key components for the stable suppression of red scale: the very rapid reproduction rate of A. melinus whose generation time is shorter than that of red scale, and the long invulnerable adult stage of female red scale. These two features reduce the time lag of the response of parasitoids to increasing host density, and reduce the potential for red scale population crash as parasitoid populations peak. The most remarkable feature of this experiment is that stable pest control is achieved at small spatial scales – single trees. A final mechanism that can stabilize the inherent tendency for host–parasitoid dynamics to cycle is immigration of hosts and/or parasitoids when they are at low densities and emigration when they are at high densities (Reeve, 1988; Godfray & Pacala, 1992; Ives, 1992a; Murdoch & Briggs, 1997). When this is the case, peak densities of either host or parasitoid may be damped


Fig. 6.2 (A) Densities of red scale (solid lines) and adult Aphytis melinus (dashed lines) averaged among four experimental trees (upper panel) on which red scale densities were experimentally increased by 80-fold, and ten control trees (lower panel). Data from Murdoch et al. (2005). (B) Pea aphid densities and percent parasitism in an alfalfa field over the course of a growing season. Pea aphid densities are given as numbers per stem, and percent parasitism was determined by dissection. Arrows mark alfalfa harvesting. Data from Gross et al. (2005).

and low population densities may be rescued. This appears to occur in the alfalfa–pea aphid–A. ervi system in the field (Gross et al., 2005), in contrast to in greenhouse cages. Alfalfa fields are harvested two to four times per growing season in Wisconsin, causing 2–3 orders of magnitude reductions in pea aphid densities (Fig. 6.2B). Following harvest, pea aphid populations rebound quickly from both the few survivors and immigrant winged aphids (Rauwald & Ives, 2001). After a few pea aphid generations (a few weeks), aphid populations often plateau or even decline. The pea aphid dynamics can look remarkably similar to those observed in the greenhouse cages (Fig. 6.1), with increases and declines occurring over a 40-day time period (Fig. 6.2B, first harvesting cycle). Nonetheless, detailed statistical analyses of these and similar data show that the plateauing and sometimes decline of

pea aphid densities may occur without increased parasitism (Gross et al., 2005), as percent parasitism remains roughly constant over the course of a harvesting cycle (Fig. 6.2B). Lack of temporal changes in parasitism with changes in aphid density is likely due to the high movement rates of A. ervi, as females move readily among fields and do not appear to show a strong response to spatial variation in pea aphid densities among fields. When parasitism does not increase with increasing pea aphid density, A. ervi do not drive population cycles like those observed in greenhouse cages, nor do they provide density-dependent control of pea aphid populations within fields. Density-dependent control of pea aphid populations may involve a variety of generalist predators that immigrate into alfalfa fields as pea aphid densities increase (Cardinale et al., 2003; Snyder & Ives, 2003). Even when A. ervi do not provide densitydependent control, however, the impact of parasitism may be high; in the example illustrated in Fig. 6.2B, on average, 20% of non-parasitized aphids were parasitized every day, and in the absence of parasitism, pea aphid population densities would likely increase to roughly ten times the observed densities over the course of a harvesting cycle (Gross et al., 2005). These examples show that, while effective biological control agents can be specialists with high




reproduction rates, stable suppression of pests likely requires attributes in addition to the three proposed by Debach & Rosen (1991). In the red scale–Aphytis melinus system, stability occurred on a very small spatial scale (single trees) due to an invulnerable adult stage of red scale. In contrast, the pea aphid–Aphidius ervi system is very “open,” with aphids and parasitoids moving readily among fields. While this prohibits A. ervi from exterminating pea aphid populations within fields, parasitism is nonetheless a major source of aphid mortality that, in combination with predation, maintains low densities of pea aphids that rarely exceed one aphid per alfalfa stem.

6.2 Generalist predators: the good, the bad and the ugly At the other end of the diet-breadth spectrum from specialists lie the generalists. Generalists feed on several or many prey species. A given pest species is often attacked by a small group of specialists, but a diverse community of generalists. The numerous natural enemies of pea aphids in Wisconsin provide an example (Snyder & Ives, 2003). The most common generalists include predatory damsel and minute-pirate bugs (Nabis spp. and Orius spp.) and ground and ladybird beetles (in the families Carabidae and Coccinellidae). The alfalfa system raises two points about generalist predators. First, while for simplicity we present a dichotomy between specialists and generalists, in reality many species blur these distinctions. Damsel bugs, for example, are broad generalists that feed upon any soft-bodied prey that they encounter and can subdue (Lattin, 1989). In contrast, the ladybird beetles common to this community, Coccinella septempunctata and Harmonia axyridis, feed primarily on aphids, albeit aphids of many different species, while also opportunistically feeding upon the eggs of other insects, pollen or other easily captured or vulnerable prey (Evans, 1991; Hodek & Honek, 1996; LaMana & Miller, 1996; Yasuda et al., 2004). Second, in the alfalfa system all of the generalists are predators in that they feed on multiple prey items during their lifetime. Some parasitoid species can also be

Fig. 6.3 Population dynamics of pea aphids on alfalfa, redrawn from data presented in Snyder & Ives (2003). Predators suppressed aphid densities (Pred) compared to controls lacking predators (O). However, aphid densities in both treatments continued to grow throughout the experiment. Uncaged reference plots (Open) followed aphid dynamics in the alfalfa field surrounding the cages.

quite diverse in their feeding habits, and in some cases their impacts on and interactions with other species may be quite like those of generalist predators (e.g. Montoya et al., 2003). However, predators that are generalists have been particularly well studied and thus we focus this section of our chapter on generalist predators. Generalist predators feed on multiple prey species and often have relatively long generation times compared to their prey. Thus, these predators do not fit two of the key attributes Debach & Rosen (1991) propose for effective biocontrol agents. Generalized feeders that reproduce slowly will be unable to mount a density-dependent population response to increasing pest densities. We found this to be the case when examining the impact of the guild of generalist predators of pea aphids in alfalfa, in the absence of the specialist A. ervi. We established cages housing alfalfa plants infested with pea aphids, with or without the diverse community of generalist predators also present, and then followed pea aphid dynamics through time. Predators reduced pea aphid densities initially compared to no-predator controls, a reduction maintained through time (Fig. 6.3). However, the subsequent aphid


population growth rate was little changed by the addition of predators, and aphid densities in both treatments continued to grow throughout the 12-day course of the experiment (Fig. 6.3). Thus, while predators caused considerable aphid mortality, they did not control aphid densities in a density-dependent fashion. In addition to having low reproduction rates and hence slow population responses to pest outbreaks, generalists sometimes attack one another rather than pests. Ecologists call predation of one natural enemy by another intraguild predation (Polis et al., 1989). When strong, intraguild predation can negate any benefits of generalist predators for biological control (Rosenheim et al., 1995). Rosenheim et al. (1993) present a classic example of intraguild predation’s disruptive effects. They examined control of cotton aphid (Aphis gossypii) on cotton by a community of generalist predators (Fig. 6.4). Key among these natural enemies were lacewing (Chrysoperla carnea) larvae, voracious predators of aphids. When alone, lacewings consumed far more aphids per capita than did any other predator species in the community. Unfortunately, the other predators were quite effective at capturing and eating lacewings, so that pairings with other predators generally led to intraguild predation of lacewings and weaker aphid control (Rosenheim et al., 1993). In the field lacewings co-occur with many other natural enemies, so aphid biological control likely will always fall below the ideal level of suppression that lacewings alone might otherwise achieve.

6.2.1 Effectiveness of biological control by generalists Despite the two disadvantages described above – low population growth and intraguild predation – generalists may have several attributes that allow them to be effective biological control agents. Although the often low population growth rate of generalists relative to their prey prohibits generalists from mounting a strong densitydependent response to increasing pest abundance through reproduction, they may nonetheless show a behavioral response, immigrating into fields with high pest densities and remaining longer to consume their prey (Frazer & Gill, 1981; Evans & Youssef, 1992). This behavioral response of

Fig. 6.4 Population growth rates for cotton aphids on cotton plants, redrawn from Rosenheim et al. (1993). Aphid densities declined when lacewing larvae (L) were alone. However, aphid populations grew when this single most effective aphid predator species was paired with Geocoris (L+G), Nabis (L+N) or Zelus (L+Z) bugs, other common predators in the cotton system and intraguild predators of lacewing larvae.

generalist predators apparently underlies the control of pea aphids in open field conditions (Fig. 6.2B). In contrast to the case in field cages, open fields allow the immigration of generalist preda¨ tors (e.g. Ostman & Ives, 2003) that in combination suppress aphid populations in a densitydependent manner. Because generalists feed on several different prey species, their densities in a crop will not be entirely tied to the density of any particular pest (Harmon & Andow, 2004). For this reason generalists may already be present in the field, feeding on non-pest prey, when pests first invade (Settle et al., 1996). Thus, generalists can serve as the “first line of defense” against pest invasion, while the relatively long life cycles of many generalist predators allow them to span peak densities of any single prey item (Symondson et al., 2002). Alternative prey are those species present in the crop other than the target pest that generalist predators will also feed upon. Alternative prey that serve to provide an additional prey resource, or that provide key nutrients otherwise limiting predator growth, can bolster predator populations and thus improve biological control (Holt, 1977).




Fig. 6.5 (A) Rice plots that did not receive pesticide applications saw an early build-up of both alternative prey and generalist predators, such that when brown planthoppers arrived predator densities were sufficient to maintain low planthopper densities. (B) In contrast, when plots received three early-season pesticide applications (indicated with arrows), build-up of alternative prey was delayed, and because predators failed to reach high densities early in the season their densities were too low to control brown planthopper later in the season. Figure depicts densities of alternative prey (OTHER), generalist predators (PRED) and brown planthopper (HERB). Note broken line in panel B; planthopper densities peaked at 1093 per m2 in this treatment. Data from Settle et al. (1996).

Settle et al. (1996) present a study demonstrating how the presence of alternative prey can build generalist predator densities to the benefit of pest suppression. These authors were working in rice crops in Indonesia, with brown planthopper (Nilaparvata lugens) as the target pest. A diverse group of generalist predators inhabits rice paddies, consisting primarily of spiders and predatory bugs. In the absence of early season insecticide application, rice crops were colonized by a diverse community of detritivores feeding on plant debris from the previous year’s crop. These detritivores served as alternative prey for the generalist predator community, allowing generalist densities to grow dramatically (Fig. 6.5A). Detritivore densities fell later in the season, just as planthoppers began to invade the crop. The generalists then

switched from attacking detritivores to attacking planthoppers, suppressing the pest below damaging levels (Fig. 6.5A). However, when insecticides were applied and detritivores were killed, generalist predator densities never built up, and planthoppers found a largely predator-free crop to colonize and reproduce in (Settle et al., 1996) (Fig. 6.5B). When predators are omnivores that include both animals and plants in their diets, non-animal foods also can have a beneficial effect on biological control. Eubanks & Denno (2000) examined biological control of pea aphids on lima beans by an omnivorous big-eyed bug (Geocoris punctipes). Big-eyed bugs feed primarily on animal prey, but will also feed on some high-quality plant parts such as bean pods. In the laboratory big-eyed bugs will reduce their feeding on aphids when in the presence of bean pods. Surprisingly though, open field plots of bean plants possessing pods exhibited aphid densities lower than those seen in plots lacking pods. Control improved in the presence of bean pods because they attracted and/or retained higher densities of big-eyed bugs, which then also fed upon aphids. Apparently, any distraction that plant feeding provided to individual predators was counteracted by the higher overall predator densities in these plots (Eubanks & Denno, 2000). With generalists nothing is simple, and the presence of alternative prey can sometimes disrupt, rather than improve, biological control. Prasad & Snyder (2006) present an example of such disruption due to the presence of alternative prey. Here, cabbage root maggot (Delia radicum) were the focal pest, and a community of surface-foraging ground and rove beetles (in the families Carabidae and Staphylinidae) were the biological control agents. Densities of ground beetles increased twofold following the installation of grassy strips into cropping fields; the beetles use the grassy strips to overwinter and escape pesticide applications and tilling. However, root maggot control was not improved following this dramatic doubling in predator density. Subsequent field and laboratory experiments revealed that aphids commonly occurred in Brassica fields and served as alternative prey. The ground predators found aphids to be tastier prey than root maggots, and when given


Fig. 6.6 The green world hypothesis envisions distinct trophic levels (A), with predators suppressing pests and in turn protecting the crop from herbivory through trophic cascades. In contrast, when intraguild predation is common (B), predators interfere with one another, disrupting the trophic cascade and helping free the pest from control. Agroecologists have proposed a third scenario (C), wherein multiple, complementary natural enemies work together to suppress pests. Figure depicts the crop (Crop), pest (Pest), primary predator(s) (Predator) and intraguild predator (IG Predator). Arrows depict energy flow and thus point from resource to consumer, and are scaled to reflect the strength of that interaction.

this choice beetles feasted on aphids and ignored root maggots (Prasad & Snyder, 2006).

6.3 Natural enemy biodiversity and biological control As growers reduce their use of broad-spectrum insecticides, the abundance and diversity of natural enemies increases (Bengtsson et al., 2005; Hole et al., 2005). Presumably a greater abundance of natural enemies is helpful for biological control, but the importance of species diversity is less clear. Here we first discuss different ways that ecologists have envisioned the relationship between natural enemy diversity and the effectiveness of biological control. Next, we discuss the positive interactions among natural enemies that impact how pest control is influenced by natural enemy biodiversity. Ecologists have advanced three hypotheses concerning the relationship between the num-

ber of predator species present and the success of biological control. Hairston et al. (1960) argued that predators can be grouped into a cohesive third trophic level, acting in concert to regulate herbivores. As predators kill herbivores they indirectly protect plants through so-called trophic cascades (Fig. 6.6A). This theory is sometimes referred to as the green world hypothesis, and it implies that the top–down impact of predators is consistently strong regardless of how many species are present (Hairston & Hairston, 1993, 1997). However, communities often include many species of generalist predators. Generalist predators usually feed on members of several trophic levels, and these catholic feeding habits might blur the distinction of discrete trophic levels, thus diffusing trophic cascades (Polis, 1991; Strong, 1992; Polis & Strong, 1996) (Fig. 6.6B). This viewpoint is known as the trophic level omnivory hypothesis, and the prediction here is that while trophic cascades may occur in some simple systems (e.g. small islands or agricultural monocultures), top– down effects will weaken as species diversity increases. Biological control theory provides a third view of the relationship between predator biodiversity and predator impact, predicting that diversifying agroecosystems, for example by intercropping, should lead to a diversified prey base and thus a more abundant and diverse predator community (Pimentel, 1961; van Emden & Williams, 1974) (Fig. 6.6C). With growing diversity predators are expected to provide more consistent prey suppression, what Root (1973) called the enemies hypothesis. Thus, frustratingly, theory offers three conflicting viewpoints predicting unchanging,




decreasing, or increasing strength of predator impact with increasing biodiversity. Worse still, recent experimental studies wherein predator diversity has been manipulated have uncovered stronger, weaker and unchanged pest suppression with greater predator biodiversity (Cardinale et al., 2006; Straub et al., 2008). How pest control varies with diversity in any particular system may reflect the balance between negative and positive interactions among constituent species (Ives et al., 2005). Negative relationships between predator biodiversity and herbivore suppression are generally attributed to intraguild predation (e.g. Finke & Denno, 2004). While less is known about the positive relationships among predators that underlie improved herbivore suppression with greater predator diversity (Ives et al., 2005), positive predator–predator interactions have begun to receive some attention, and we next review the beneficial effects of greater natural enemy biodiversity.

6.3.1 Natural enemies that complement one another It is thought that adding species to a community can improve biological control when those predators differ in where, how or when they attack the pest, such that each predator fills a unique role (Wilby & Thomas, 2002). When this occurs, multiple enemy species can complement one another, with a sufficiently diverse predator community leaving herbivores with no safe refuge in time or space (Casula et al., 2006). Further, predators that use different microhabitats are less likely to encounter and interfere with one another than are predators that use the same microhabitat (Schmitz, 2005). Wilby et al. (2005) provide a nice example of how complementary relationships among predators may develop, working with two pests of rice, rice leaf-folder (Marasmia patnalis, a moth) and brown planthopper. Control was compared among single versus three-species assemblages of natural enemies, chosen from among a group of predators including a wolf spider (Pardosa pseudoannulata), a ladybird beetle (Micraspis crocea), a predatory cricket (Metioche vittaticollis) and a plant bug (Cyrtorhinus lividipennis). Predator densities were scaled across all treatments such that

predation pressure was equalized, meaning that any differences between single and multi-species treatments would result from an emergent benefit of species diversity per se. Biological control improved with greater predator diversity for the moth, but not for the planthopper. The moth had a complex life cycle with stages that differed dramatically in morphology and behavior, whereas the leafhopper had a simple life cycle with successive stages closely resembling one another. The authors conclude that predator diversity was important for moth control because different predators attacked the different moth stages; i.e. the predators complemented one another. For the planthopper, in contrast, all predators impacted the different pest stages similarly, such that natural enemies had identical, redundant effects and thus did not complement one another (Wilby et al., 2005). Snyder & Ives (2003) examined complementary interactions of a different sort among natural enemies. We compared biological control by the specialist parasitoid A. ervi to that exerted by a diverse community of generalist predators within field cages (Fig. 6.7). The parasitoid A. ervi on its own was capable of mounting a strong numerical response to aphid increase, but this occurred at a time lag such that aphids reached a relatively high peak density before eventually declining. The generalist predators exerted significant suppression of aphids early on, but their impact did not increase as aphid densities grew and high peak aphid densities again were reached. However, when both the parasitoid and predators were present together, resulting biological control combined the beneficial attributes of both enemy classes, with both early suppression of aphids by predators and later suppression by the parasitoid, thereby leading to lower aphid densities throughout the experiment (Snyder & Ives, 2003). Thus, the densityindependent impact of predators complemented the density-dependent response of the parasitoid.

6.3.2 Predators that facilitate one another’s prey capture A second mechanism thought to contribute to improved pest suppression with greater natural enemy diversity is predator–predator


way the combined impact of ladybirds and ground beetles on aphids, when both species were present simultaneously, was greater than could be predicted based on the effectiveness of either species in isolation (Losey & Denno, 1998).

6.4 Summary

Fig. 6.7 Population dynamics of pea aphid on alfalfa, redrawn from data presented in Snyder & Ives (2003). Peak aphid densities were highest in controls that included no natural enemies (O). When alone the parasitoid A. ervi (Para) had little impact on aphid population growth early in the experiment, but drove down aphid densities after day 14. The diverse community of generalist predators alone (Pred) immediately reduced aphid densities, but thereafter aphid densities generally ran parallel to those in no-enemy controls. Aphid suppression was most effective with both parasitoid and predators present (Both), because the early suppression exerted by generalists was combined with later control by parasitoids. In this way parasitoid and predators complemented one another. Uncaged reference plots (Open) followed aphid dynamics in the alfalfa field surrounding the cages.

facilitation. Facilitation occurs when predators indirectly enhance one another’s success in prey capture. This occurs, for example, when prey leave a habitat to escape one predator, only to instead fall victim to a second predator species in the would-be refuge (Sih et al., 1998). In these cases prey find themselves “between a rock and hard place,” with one predator species chasing prey into the waiting jaws of a second predator. In one well-known study from the alfalfa–pea aphid system (Losey & Denno, 1998), aphids dropped to the ground to escape ladybird beetles (Coccinella septempunctata) foraging in the foliage. However, once on the ground the aphids were then vulnerable to attack by ground beetles (e.g. Harpalus pennsylvanicus) foraging on the soil surface; in the absence of ladybird beetles to chase aphids to the soil the ground beetles rarely encountered aphids. In this

Specialists attack a narrow range of pest species and often reproduce at a rate greater than that of their prey. These traits lead to specialists’ ability to mount a density-dependent population response to rising prey densities, thought to be a trait important for an effective biological control agent. However, these same attributes also increase the chances that specialists drive boomand-bust cycles, rather than providing stable control of prey densities. Generalists generally lack the ability to mount a density-dependent population response but are able to switch to attacking pests as they become abundant, either by immigrating into fields with pest outbreaks or switching from alternative prey within the same fields. Thus, both specialists and generalists have advantages and limitations as biological control agents. Different theories about the suppression of herbivores by their natural enemies predict stronger, weaker or unchanged pest suppression with greater natural-enemy diversity. How predator biodiversity influences pest control likely reflects a balance between positive and negative predator–predator interactions. Intraguild predation can disrupt biological control, and when intraguild predation is common pest suppression is expected to grow weaker when more natural enemy species are present. In contrast, when predators complement one another by attacking the pest in unique ways, or facilitate one another’s prey capture, then pest suppression is expected to improve when more natural enemy species are present.

Acknowledgments W. E. Snyder was supported during preparation of this chapter by grant #2004–01215 from the National Research Initiative of the US Department




of Agriculture (USDA), Cooperative Research, Education and Extension Service (CSREES). Much of the work on pea aphids was supported by grants from the USDA and National Science Foundation (NSF) to A. R. Ives and by a USDA–National Research Initiative (NRI) Postdoctoral Fellowship to W. E. Snyder.

References Bengtsson, J., Ahnstr¨ om, J. & Weibull, A. (2005). The effects of organic agriculture on biodiversity and abundance: a meta-analysis. Journal of Applied Ecology, 42, 261–269. Briggs, C. J. & Hoopes, M. F. (2004). Stabilizing effects in spatial parasitoid–host and predator–prey models: a review. Theoretical Population Biology, 65, 299– 315. Cardinale, B. J., Harvey, C. T., Gross, K. & Ives, A. R. (2003). Biodiversity and biocontrol: emergent impacts of a multi-enemy assemblage on pest suppression and crop yield in an agroecosystem. Ecology Letters, 6, 857– 865. Cardinale, B. J., Srivastava, D., Duffy, J. E. et al. (2006). Effects of biodiversity on the functioning of trophic groups and ecosystems. Nature, 443, 989– 992. Casula, P., Wilby, A. & Thomas, M. B. (2006). Understanding biodiversity effects on prey in multi-enemy systems. Ecology Letters, 9, 995–1004. Debach, P. & Rosen, D. (1991). Biological Control by Natural Enemies. New York: Cambridge University Press. Denoth, M., Frid, L. & Myers, J. H. (2002). Multiple agents in biological control: improving the odds? Biological Control, 24, 20–30. Eubanks, M. D. & Denno, R. F. (2000). Host plants mediate ominivore–herbivore interactions and influence prey suppression. Ecology, 81, 936–947. Evans, E. W. (1991). Intra versus interspecific interactions of ladybeetles (Coleoptera: Coccinellidae) attacking aphids. Oecologia, 87, 401–408. Evans, E. W. & Youssef, N. N. (1992). Numerical responses of aphid predators to varying prey density among Utah alfalfa fields. Journal of the Kansas Entomological Society, 65, 30–38. Finke, D. L. & Denno, R. F. (2004). Predator diversity dampens trophic cascades. Nature, 429, 407– 410. Frazer, B. D. & Gill, B. (1981). Hunger, movement and predation of Coccinella californica on pea aphids in the laboratory and in the field. Canadian Entomologist, 113, 1025–1033.

Godfray, H. C. J. & Pacala, S. W. (1992). Aggregation and the population dynamics of parasitoids and predators. American Naturalist, 140, 30–40. Gross, K. & Ives, A. R. (1999). Inferring host–parasitoid stability from patterns of parasitism among patches. American Naturalist, 154, 489–496. Gross, K., Ives, A. R. & Nordheim, E. V. (2005). Estimating time-varying vital rates from observation time series: a case study in aphid biological control. Ecology, 86, 740–752. Hairston, N. G. Jr. & Hairston, N. G. Sr. (1993). Cause– effect relationships in energy flow, trophic structure, and interspecies interactions. American Naturalist, 142, 379–411. Hairston, N. G. Jr. & Hairston, N. G. Sr. (1997). Does food-web complexity eliminate trophic-level dynamics? American Naturalist, 149, 1001–1007. Hairston, N. G., Smith, F. E. & Slobodkin, L. B. (1960). Community structure, population control and competition. American Naturalist, 94, 421–425. Harmon, J. P. & Andow, D. A. (2004). Indirect effects between shared prey: predictions for biological control. BioControl, 49, 605–626. Hassell, M. P. (1978). The Dynamics of Arthropod Predator– Prey Systems. Princeton, NJ: Princeton University Press. Hassell, M. P. & May, R. M. (1973). Stability in insect host– parasite models. Journal of Animal Ecology, 42, 693–736. Hassell, M. P. & May, R. M. (1974). Aggregation in predators and insect parasites and its effect on stability. Journal of Animal Ecology, 43, 567–594. Hassell, M. P., May, R. M., Pacala, S. W. & Chesson, P. L. (1991). The persistence of host–parasitoid associations in patchy environments. I. A general criterion. American Naturalist, 138, 568–583. Hawkins, B.A., Mills, N. J., Jervis, M. A. & Price, P. W. (1999). Is the biological control of insects a natural phenomenon? Oikos, 86, 493–506. Hodek, I. & Honek, A. (1996). Ecology of Coccinellidae. Boston, MA: Kluwer. Hole, D. G., Perkins, A. J., Wilson, J. D. et al. (2005). Does organic farming benefit biodiversity? Biological Conservation, 122, 113–139. Holt, R. D. (1977). Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology, 12, 197–229. Huffaker, C. B. (1958). Experimental studies on predation: dispersion factors and predator–prey oscillations. Hilgardia, 27, 343–383. Ives, A. R. (1992a). Continuous-time models of host– parasitoid interactions. American Naturalist, 140, 1–29. Ives, A. R. (1992b). Density-dependent and densityindependent parasitoid aggregation in model


host–parasitoid systems. American Naturalist, 140, 912– 937. Ives, A. R. (1995). Spatial heterogeneity and host– parasitoid population dynamics: do we need to study behavior? Oikos, 74, 366–376. Ives, A. R., Cardinale, B. J. & Snyder, W. E. (2005). A synthesis of subdisciplines: predator–prey interactions, and biodiversity and ecosystem functioning. Ecology Letters, 8, 102–116. LaMana, M. L. & Miller, J. C. (1996). Field observations on Harmonia axyridis Pallas (Coleoptera: Coccinellidae) in Oregon. Biological Control, 6, 232–237. Landis, D. A., Wratten, S. D. & Gurr, G. M. (2000). Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175–201. Lattin, J. D. (1989). Bionomics of the Nabidae. Annual Review of Entomology, 34, 383–400. Losey, J. E. & Denno, R. F. (1998). Positive predator– predator interactions: enhanced predation rates and synergistic suppression of aphid populations. Ecology, 79, 2143–2152. Mackauer, M. & Kambhampati, S. (1986). Structural changes in the parasite guild attacking the pea aphid in North America. In Ecology of Aphidophaga, ed. I. Hodek, pp. 347–356. Prague, Czechoslovakia: Academia. Montoya J. M., Rodr´ıguez, M. A. & Hawkins, B. A. (2003). Food web complexity and higher-level ecosystem services. Ecology Letters, 6, 587–593. Murdoch, W. W. & Briggs, C. J. (1997). Theory for biological control: recent developments. Ecology, 77, 2001– 2013. Murdoch, W. W., Nisbet, R. M., Luck, R. F., Godfray, H. C. J. & Gurney, W. S. C. (1992). Size-selective sex-allocation and host feeding in a parasitoid host model. Journal of Animal Ecology, 61, 533–541. Murdoch, W. W., Briggs, C. J. & Nisbet, R. M. (2003). Consumer-Resource Dynamics. Princeton, NJ: Princeton University Press. Murdoch, W., Briggs, C. J. & Swarbrick, S. (2005). Host suppression and stability in a parasitoid–host system: experimental demonstration. Science, 309, 610– 613. Myers, J. H., Higgins, C. & Kovacs, E. (1989). How many insect species are necessary for the biological control of insects? Environmental Entomology, 18, 541–547. Nicholson, A. J. & Bailey, V. A. (1935). The balance of animal populations. Proceedings of the Zoological Society of London, 43, 551–598. Olson, A. C., Ives, A. R. & Gross, K. (2000). Spatially aggregated parasitism on pea aphids, Acyrthosiphon pisum,

caused by random foraging behavior of the parasitoid Aphidius ervi. Oikos, 91, 66–76. ¨ ¨ & Ives, A. R. (2003). Scale-dependent indiOstman, O. rect interactions between two prey species through a shared predator. Oikos, 102, 505–514. Pimentel, D. (1961). Species diversity and insect population outbreaks. Annals of the Entomological Society of America, 54, 76–86. Polis, G. A. (1991). Complex trophic interactions in deserts: an empirical critique of food-web theory. American Naturalist, 138, 123–155. Polis, G. A. & Strong, D. R. (1996). Food web complexity and community dynamics. American Naturalist, 147, 813–846. Polis, G. A., Myers, C. A. & Holt, R. D. (1989). The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology and Systematics, 20, 297–330. Prasad, R. P. & Snyder, W. E. (2006). Polyphagy complicates conservation biological control that targets generalist predators. Journal of Applied Ecology, 43, 343–352. Rauwald, K. S. & Ives, A. R. (2001). Biological control in disturbed agricultural systems and the rapid reestablishment of parasitoids. Ecological Applications, 11, 1224–1234. Reeve, J. D. (1988). Environmental variability, migration and persistence in host–parasitoid systems. American Naturalist, 132, 810–836. Root, R. B. (1973). Organization of a plant–arthropod association in simple and diverse habitats: the fauna of collards. Ecological Monographs, 43, 95–124. Rosenheim, J. A., Wilhoit, L. R. & Armer, C. A. (1993). Influence of intraguild predation among generalist insect predators on the suppression of an herbivore population. Oecologia, 96, 439–449. Rosenheim, J. A., Kaya, H. K., Ehler, L. E., Marois, J. J. & Jaffee, B. A. (1995). Intraguild predation among biological-control agents: theory and practice. Biological Control, 5, 303–335. Schmitz, O. J. (2005). Behaviors of predators and prey and links with population-level processes. In Ecology of Predator–Prey Interactions, eds. P. Barbosa & I. Castellanos, pp. 256–278. New York: Oxford University Press. Settle, W. H., Ariawan, H., Astuti, E. T. et al. (1996). Managing tropical rice pests through conservation of generalist natural enemies and alternative prey. Ecology, 77, 1975–1988. Sih, A., Englund, G. & Wooster, D. (1998). Emergent impacts of multiple predators on prey. Trends in Ecology and Evolution, 13, 350–355. Snyder, W. E. & Ives, A. R. (2003). Interactions between specialist and generalist natural enemies: parasitoids,




predators, and pea aphid biocontrol. Ecology, 84, 91– 107. Straub, C. S., Finke, D. L. & Snyder, W. E. (2008) Are the conservation of natural enemy biodiversity and biological control compatible goals? Biological Control, 45, 225–237. Strong, D. R. (1992). Are trophic cascades all wet? Differentiation and donor-control in speciose systems. Ecology, 73, 747–754. Symondson, W. O. C., Sunderland, K. D. & Greenstone, M. H. (2002). Can generalist predators be effective biocontrol agents? Annual Review of Entomology, 47, 561– 594. Tscharntke, T., Klein, A., M. Kruess, A., Steffan-Dewenter, I. & Thies, C. (2005). Landscape perspectives on agricultural intensification and biodiversity – ecosystem service management. Ecology Letters, 8, 857– 874.

Turchin, P. (2003). Complex Population Dynamics: A Theoretical/Empirical Synthesis. Princeton, NJ: Princeton University Press. van Emden, H. F. & Williams, G. F. (1974). Insect stability and diversity in agro-ecosystems. Annual Review of Entomology, 19, 455–475. Wilby, A. & Thomas, M. B. (2002). Natural enemy diversity and pest control: patterns of pest emergence with agricultural intensification. Ecology Letters, 45, 353– 360. Wilby, A., Villareal, S. C., Lan, L. P., Heong, K. L. & Thomas, M. B. (2005). Functional benefits of predator species diversity depend on prey identity. Ecological Entomology, 30, 497–501. Yasuda, H., Evans, E. W., Kajita, Y., Urakawa, K. & Takizawa, T. (2004). Asymmetric larval interactions between introduced and indigenous ladybirds in North America. Oecologia, 141, 722–731.

Chapter 7

Sampling for detection, estimation and IPM decision making Roger D. Moon and L. T. (Ted) Wilson IPM requires information about populations of pest and beneficial organisms in managed and constructed habitats. A recurring question is whether potentially injurious pests are abundant enough to warrant intervention, or whether beneficial organisms or other factors are likely to maintain control. Because most managed habitats are too large to be examined completely, practitioners must sample them and draw an inference about the whole. Sampling plans in IPM can be grouped into three categories, depending on the sampler’s goal. First, detection sampling is used in surveillance and regulatory applications, where the critical density of a target organism is effectively zero. Detection plans are designed to control the chance that the organism is erroneously missed. Second, estimation sampling is used where the goal is to quantify abundance, usually with desired levels of precision and reliability. Estimation sampling is used mainly in research, but it can also be used to evaluate IPM implementation and effectiveness. Finally, decision sampling is used where a choice to intervene with one or more management tactics depends on whether abundance has or will soon exceed a threshold density. Rather than estimate density, the goal is more simply to classify the habitat as needing or not needing intervention.

In all three situations, the basic process is the same. A sampler selects a set of sample units from the habitat using a defined procedure, assesses each for presence or abundance of the target organism, and then draws a conclusion based on the results. Fundamental sources on sampling theory and survey design are Cochran (1977) and Lohr (1999). Kogan & Herzog (1980), Pedigo & Buntin (1994) and Southwood & Henderson (2000) review sampling techniques used in ecology and agriculture. Binns et al. (2000) review theory and techniques for design and analysis of sampling plans for decision making in crop IPM. Sampling has elements of art and science, but the underlying foundation is mathematical. The goals of this chapter are to provide an overview of sampling theory and design techniques that are the basis of practical IPM sampling.

7.1 Basics of sampling Sampling of all three kinds begins with a clear understanding of the attribute to be measured. In IPM, pest abundance is commonly indexed by the proportion of units that are infested, or as mean density (number of individuals) per sample unit. Of course, other variables may be of interest,

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



too, depending on management context. Next, the sample universe must be defined. It is usually a specific habitat with limits in space and time. The sample unit is the basic entity that is actually observed. A set of units constitutes a sample. To be used operationally, sample units should be mutually exclusive, convenient to locate and easy to process. A general recommendation is to choose units that are physically small, because more can be processed affordably, and the resulting estimates tend to be more precise than if larger units are used. Sample units can be direct, physical subdivisions of a habitat, such as soil cores or quadrats of known area. These provide measures of abundance per unit of habitat area. If a universe of area U is sampled with a unit of area a, then there will be U/a = N possible units in the universe. In contrast, units may have dimensions that are more ambiguous, as is the case with leaves, whole plants, or aerial collection devices. One problem with these kinds of sample units is that N may be difficult to calculate because unit size is unknown. This problem will be unimportant as long as unit size and number can be assumed to remain constant. However, in cases where size, number and occupancy by target organisms might change through time, then results will need to be interpreted with caution. Estimates of abundance from one kind of sample unit can be converted to another kind if prior research has established a basis for calibration. To develop notation to be used in this chapter, the sampler inspects i = 1 . . . n of the N units, and records a series of xi observations. The ratio f = n/N is known as the sampling fraction, and in most cases will be trivially small. Sampling is almost always done without replacement, meaning individual sample units are observed only once. As long as N > ∼500, reliability of a sample is determined by the number of units examined (n), and not by the fraction examined ( f ).

7.2 Detection sampling This form of sampling is used to test the hypothesis that a pest is absent in a habitat. Examples include trapping to map geographic limits of

invading forest pests, sampling to certify absence of disease in seed lots and inspection of cargo containers at ports of entry to exclude exotic fruit or vegetable insects. Sample universes and sample units in these situations can range from a large number (N >100 000) of possible trap locations in a region, down to a relatively small number (N ∼ 100) of cartons in a shipping container. Detection sampling is done sequentially, where individual units are examined one at a time, beginning with units most likely to be infested. When the first infested one is found, sampling stops and the habitat is declared to be infested. In contrast, as long as all units are clean, sampling could continue until all N units are examined, which will be impractical if N is large. Absence can never be proven, but sampling can be designed to control the chance of detecting an infestation when proportion infested is greater than a specified cut-off level, pc (Venette et al., 2002).

7.2.1 Necessary sample size A design problem is to determine how many pest-free units one would need to conclude with desired reliability that the true proportion infested is not above the cut-off, pc . Statistical theory indicates that if a sample of n units is taken without replacement from a universe of size N, and each unit is tested with a procedure that is 100% sensitive and specific, then the probability of observing no infested units can be calculated as the “zero-term” of a hypergeometric distribution (Kuno, 1991). The number of non-infested units, n0 , needed to conclude that the true proportion is not above the cut-off level is n0 = N (1 − α 1/N pc ).


The term α is the probability of missing an infested unit when one is actually present, and is customarily set at 0.05, but can be liberal, 0.2 or greater, or conservative, 0.01 or lower. That probability translates into a level of reliability, which is the 100(1 – α)% of sampling occasions that can be expected to detect an infestation. Values for pc and α are chosen when a detection plan is designed, and will be based on an economic balance between cost of sampling and potential losses of failing to detect a pest when it is actually present.


absence would have a reliability of 100(1 – 0.904)%, just under 10%. Increasing n0 to 100 would raise reliability to 65%. Retrospective sampling to determine the probability of a failure of detection is used by the State of Minnesota, Department of Transportation (MN-DOT) to monitor abundance of the roadside weed leafy spurge (Euphorbia esula) (Moon, 2007b).

7.3 Estimation sampling

Fig. 7.1 Number of non-infested sample units (n0 ) needed to declare with two levels of reliability that true infestation rate is below three chosen cut-off levels (pc ), over a range of universe sizes, via Eq. (7.1). Note n0 and N in log10 scale.

Solutions of Eq. (7.1) with arbitrary cut-off and reliability levels over a range of Ns are illustrated in Fig. 7.1. In small universes (relatively low N), virtually all units will need to be examined, regardless of cut-off level, to be 80% or more certain that abundance is not above the cut-off level. In larger universes, sample size depends mainly on chosen cut-off level and reliability. If pc is set at a liberal 1% (1 in 100 units infested), then over the range of N, sample sizes needed to achieve 80% or 95% reliability reach upper asymptotes of 160 or 300, respectively. If pc is a more conservative 0.01% (1 in 10 000), then maxima of ∼16 000 or 30 000 will be required.

7.2.2 Chance of detection A complementary problem is to evaluate retrospectively the chance that an infestation would have been missed, given that a fixed number of non-infested units had been observed. Rearrangement of Eq. (7.1) leads to  n0  N pc α∗ = 1 − , N


where α ∗ is the calculated chance that an infestation would have been missed with a sample size of n0 . To illustrate, if a universe contained N = 1000 units, the cut-off level was set at 1% infested, and n0 = 10 units were found to be non-infested, then the chance of missing the infestation would be 0.904. Stated another way, the declaration of

This type of sampling is designed to measure the level of a trait of interest. Equations for calculating sample statistics for common IPM variables are summarized in Table 7.1. Relevant statistics are the sample mean (¯ x , an estimate of true mean, µ), the sample variance (s2 , an estimate of σ 2 ) and the standard error of the mean (SE). The term (1 – f) in the formulas for SE adjusts downward for sampling without replacement from a finite population. The estimators in Table 7.1 are appropriate for simple random sampling, where every sample unit has an equal chance of being observed. This sampling design and others will be discussed later.

7.3.1 Precision and reliability of estimates Samplers can control both the precision and reliability of estimated means. Precision can be represented in absolute terms by h = Student’s t × SE, which is in the same measurement units as the mean. Alternatively, precision can be expressed in relative terms, as a unitless fraction (or percentage) of the mean, d = Student’s t × SE/¯ x . Either way, precision is calculated as a confidence interval for an estimated mean, CI = x¯ ± h or x¯ ± d¯ x: the narrower the confidence interval, the more precise the estimate. Reliability of a confidence interval is governed by Student’s t, which is chosen in accord with permissible type I error rate (α). In estimation sampling, a type I error occurs when a confidence interval fails to contain the true mean by chance alone. Convention is for samplers to set α at 0.05 or 0.10, allowing true means to fall outside an interval in 1/20 or 1/10 sampling events, respectively. For n > 30, values of Student’s t will be ∼1.0, 1.7 or 2.0 for α = 0.33, 0.10 and 0.05, respectively.




Table 7.1

Equations for estimating sample statistics for variables commonly measured in IPM

Variable of interest


Continuous, amount per unit, (normal distribution)

x¯ =

1 n

Proportion infested (binomial)

pˆ =

1 n

Density, count per unit, random (Poisson)

λˆ =

1 n

n  i =1 n 

Variance n 

s =


s 2 = pˆ (1 − pˆ )

i =1 n  i =1

(x i −¯x )2



i =1




s2 =

ˆ 2 (x i −λ)

i =1


Standard errora SE = SE ≈


x¯ =

1 n

Density, count per unit (Taylor’s power relation)

x¯ =

1 n

a b

n  i =1 n  i =1

−f )

s2 (1 n−1

−f )+

1 , 2n

if n pˆ > 15, elseb  2 SE = sn (1 − f ) if n  ≥ 30, elseb λˆ (1 n

SE =

Density, count per unit, aggregated (negative binomial)

s2 (1 n


ˆ s 2 = x¯ (¯x + k)


s 2 = A x¯ b

−f ) ˆ if n λ ≥ 30, elseb  2 SE = sn (1 − f ) SE =

s2 (1 n

−f )

Term (1 − f ) adjusts for finite population, where f = n/N is fraction of available units actually examined. Exact values obtained from statistical tables, or iteratively with computer software.

The Central Limit theorem states that regardless of the underlying sampling distribution, as long as the variance is finite, sample means will tend toward being normal as sample size (n) increases. This theorem is the basis of faith that actual error rates will equal the nominal α, as long as sample size is adequate. The estimators in Table 7.1 are statistically unbiased in the sense that their values will on average equal the true but unknown values in the sample universe (Cochran, 1977). This does not mean, however, that a single estimate can not be badly biased – low or high – due to chance alone. Of greater concern, estimates can be biased consistently if there is systematic error in sample collection or measurement technique, or unequal availability of organisms due to adverse weather or differences in location due to behavior or life history.

7.3.2 Sampling distributions Four sampling distributions are commonly used to design sampling plans and to calculate associated statistics.

Normal distribution Measured variables can be continuous (real numbers) as with length, area, volume or mass. In these cases, the response variable can be modeled as having come from the familiar bell-shaped normal distribution, and estimation of a universe’s mean (µ) or total (N × µ) is straightforward (Table 7.1). In cases where the underlying distribution is skewed to the right by an excess of high values, the distribution can be modeled by a lognormal distribution (not illustrated). Binomial distribution In many IPM situations, sample units can be scored simply for presence or absence of a target organism. In this case, each xi will have a value of 1 or 0, respectively. Providing the universe is large (N effectively infinite), and the probability (p) of being positive remains constant throughout the universe, then the response can be modeled as a binomial variable with parameters p and n (Table 7.1). A property of the binomial distribution is that h for pˆ will be greatest when pˆ is 0.5, but will get smaller as pˆ approaches 0 or 1.0. Universes with clustering or a spatially variable pattern in p


can be modeled with a beta-binomial distribution (see Madden & Hughes, 1999; Binns et al., 2000). Random distribution Most frequently in IPM sampling, target organisms are counted per sample unit, and this is often referred to as enumerative sampling. Counts can be modeled as a random (Poisson) process, as might occur when fungal spores land on sticky slides or immigrating aphids land in pan traps. A random distribution arises if the organisms act independently of each other, and all sample units are equally likely to be occupied. A property of the Poisson distribution is that its mean is equal to its variance. Consequently, SE and h will be proporˆ (Table 7.1). tional to the mean, λ Aggregation In many cases, samples where organisms are counted have more low and high counts than expected from a Poisson distribution. This is a result of spatial or temporal clustering of organisms among sample units. Previous studies have shown that the large majority of species have an aggregated spatial pattern (e.g. Taylor et al., 1978; Wilson & Room, 1983). Aggregation is usually a result of unequal habitat quality and behavioral aspects of a species, but also may be affected by choice of sample unit. For examples, armyworms and stink bugs oviposit their eggs in masses. When sampling individual leaves on host plants, each leaf will contain either zero or several dozen to a few hundred eggs. At the level of a leaf, these species will be highly aggregated. After eggs hatch, mortality and dispersal will reduce aggregation. Aggregation can be modeled statistically in different ways, but a common approach is to use the negative binomial distribution (NBD), which has a mean and a second parameter k that is inversely related to extent of aggregation. Methods for estimating k and fitting alternative distributions can be found in Pedigo & Buntin (1994) and Binns et al. (2000). Generally, sample sizes must be well in excess of n = 50 to obtain reliable estimates of k, and to discriminate among alternative distributions. Experience with a variety of universes and target organisms indicates that values of k can change substantially as densities change (see

Southwood & Henderson, 2000). Furthermore, values of k and other measures of aggregation are sensitive to size of the sample unit. Aggregation tends to be greatest where sample unit size coincides with the spatial scale of an organism’s clusters. Hence, samplers can overcome aggregation (and thereby reduce h) by changing the physical size of the sample unit. Taylor’s variance–mean relation An alternative to adopting a specific statistical model for a sampling distribution is to make use of the empirical relation between observed sample variances and means among multiple samples that span a range in density. Taylor (1961) observed that sample variances and means were related as a power function (Table 7.1). By applying logs to both sides of the power function,   log10 si2 = a + b log10 (¯ xi ),


variance–mean relations among many data sets can be analyzed with simple linear regression (or more sophisticated methods: see Perry, 1981). An example is in Fig. 7.2A. The intercept, a = log10 (A), depends on sample unit size. Slope (b) is interpreted as an index of aggregation. If a is not significantly different from zero, then a test of the hypothesis that slope = 1.0 formally compares the observed variance–mean pattern with random (Poisson) expectations. Many studies indicate that values for a and b for individual species are reasonably stable across different universes, providing the same sample unit has been used. Relation between binomial distribution and density The binomial distribution is useful for describing the relation between proportion (p) of sample units infested by a species and the mean density (µ) of the species. As density increases, so does the proportion infested, and the rate at which p increases depends on extent of aggregation (Fig. 7.2B). For species that are uniformly distributed among sample units, σ 2 /µ < 1.0, p increases more rapidly with increasing density. In contrast, for species with an aggregated spatial pattern, σ 2 /µ > 1.0, p increases more slowly.




relation (Wilson & Room, 1983), p = 1 − exp(−¯ x × loge [a x¯b−1 ]/[a x¯b−1 − 1]). (7.4b)

7.3.3 Sample size

Fig. 7.2 Evidence for moderate aggregation by vegetable leafhopper (Austroasca viridigrisea) in cotton fields sampled over three growing seasons in Australia (see Wilson & Room, 1983). Each point represents a sample of vegetable leafhopper, counted visually on individual plants, n = 96 plants per date, for total of 474 samples and 45 504 plants. (A) Taylor’s variance–mean relation, log-transformed sample variances and means; dashed line random expectations, solid line least squares fit to the data (Eq. 7.3). (B) Same data expressed as proportion of plants in each sample that were infested; lines as above.

Besides setting α, samplers can control precision mainly through choice of sample size (n), which governs numerical values of both Student’s t (through degrees of freedom, df) and SE (see Table 7.1). Increasing sample size from small to moderate can substantially improve precision, but the relationship is one of diminishing returns. To illustrate, if s2 for a normal variable is 100, and α is set to 0.05, doubling n from 5 to 10 decreases h = t × SE from 12.4 to 7.2, and doubling again to n = 20 decreases it to 4.7. Providing that cost of sampling is not limiting, sample variance will ultimately limit achievable precision. Further refinement will require that sample variance be reduced, perhaps by changing the definition of the sample unit, by stratifying to direct more effort to sources of variability, or both. The simplest approach to setting sample size is to calculate required sample size (RSS), the sample size that would be needed to achieve desired precision. This procedure is useful when a sampling plan is being developed. Using chosen values for α, and either h = t × SE or d = t × SE/¯ x, RSS =

s2 s2 , or = 2 2 . 2 h d x¯

(7.5a, b)

From these general formulae, Karandinos (1976) and Ruesink (1980) derived RSS formulas for binomial, random and aggregated sampling distributions, by substituting corresponding definitions of SE (Table 7.1) into definitions of h or d. Four points are notable. First, RSS will be proportional to sample variance and inversely proportional to desired precision. The more variable the universe and the The relation between proportion infested (p) greater the precision desired, the larger the samand mean density (¯ x ) among multiple samples can ple size needed to achieve a chosen level of be expressed generally as precision. Second, RSS can change substantially, dependp = 1 − exp(−¯ x × loge [s 2 /¯ x ]/[s 2 /¯ x − 1]), (7.4a) ing on how precision is defined (as h or d) (Fig. 7.3, A vs. B). Small samples may suffice when means are where “exp” means exponentiation of natural low and precision is a positive number (h). Howbase e. The same relation can be expressed with ever, at the same density, RSS can be prohibitively variance estimated from Taylor’s mean–variance large if relative precision (d) is desired.


over what would be required if the organisms are distributed randomly among sample units. Finally, whichever way precision is specified, RSS will be proportional to sample variance, and a preliminary estimate of sample variance is required. Adequacy of calculated sample size will depend largely on accuracy of the estimated variance. In habitats where the variance is not changing rapidly, and sample units can be processed quickly, then Stein’s two-stage approach can be used. One takes a small preliminary sample to estimate variance, calculates RSS, and then quickly returns in a second stage to obtain the remaining sample units. Where the same habitat is being sampled over a series of dates, then results from one date can serve to project a value for variance on the next date.

7.3.4 Sequential sampling for desired precision

Fig. 7.3 Number of sample units required (RSS) to estimate a mean density with desired precision. (A) Desired precision defined as positive number (h). (B) Precision as proportion of the mean (d). Curves calculated with Eqs. (7.5a, b), with α = 0.05. Constant variance, s2 = 25. Random (Poisson), variance = mean. Aggregated, using Taylor’s variance–mean relation, A = 1, b = 1.5.

Third, where sample variance increases with density, as will occur when counted organisms are random or aggregated, RSS to achieve absolute precision (h) will increase as density increases, but it will decrease if relative precision (d) is desired (Fig. 7.3). Furthermore, for a given mean density, a consequence of aggregation is to increase RSS

An extension of Stein’s approach is possible in situations where sample units can be processed while sampling is in progress. The basic idea is to use the information in the growing sample to terminate sampling just when desired precision has been achieved, and to avoid excessive effort and expense. This approach is feasible where sample units can be processed quickly, and results analyzed statistically to determine when desired precision has been achieved. In cases where units must be evaluated in an off-site laboratory, then a compromise approach is to calculate RSS as above, to collect a few more units than are expected to be needed, but to process the units sequentially and thereby limit excess processing expense. Before hand-held calculators became available, much research was done to develop sequential sampling plans for use in IPM sampling. Known as stop rules, guidelines for terminating sampling were coded either in tables or graphs that conveyed a critical total of counts in relation to increasing n. A scout would begin with a total of 0, and then add additional counts as subsequent units were examined. Sampling would stop when the accumulating total first exceeded the critical stop value for the corresponding n. Equations for stop lines are based on chosen levels of precision, confidence, and sampling distribution, and have been reviewed by Hutchison (1994). We will return




below to discuss sequential procedures in decision sampling, where they are more widely adopted.

Ci =

7.3.5 Choice of sample units Size of sample unit can affect sampling cost and achievable precision. To illustrate, a sample unit might be defined as a single leaf, a cluster of three leaves, or a whole plant. Smaller units tend to be cheaper, because once located, they require less labor and materials to process. Also, depending on extent and spatial scale of aggregation, units of different sizes may yield different variances. To compare units of different size (u), Cochran (1977) defined relative net precision as RNPu =

Mu Mu × 2 , Cu su


where Mu is sample unit size (mass), Cu is cost of processing a unit, and s2 u is sample variance. The size with greatest RNP will be most efficient in operational use. Costs and variances can be estimated from a sampling experiment designed to compare units of different sizes. For some pests, qualitatively different sample units might be appropriate. For example, mobile predators can be sampled with a sweep net or by visual examination of plant parts. Which one should be used? The better technique can be chosen based on their relative costs to achieve the same sample precision (Wilson, 1994), presuming results with the different methods can be expressed in the same unit or measurement. Cost to achieve desired precision with a given method (i) can be estimated as C i = ni (θi + φi ),

proposed that each technique’s cost could be weighted by its relative acceptability,


where Ci is cost in time (or money), ni is number of units required to achieve desired precision and confidence, and θi and φi are times required to locate and process an individual unit, respectively. Sample size can be estimated as RSS (Eq. 7.5a or b) for estimation, or maximum average sample number (ASN) for decision sampling (see below). The technique with the lowest Ci would be the one of choice. To compensate for different levels of acceptability by commercial scouts, Espino et al. (2007)

ni (θi + φi ) , ψi


where ψi is the estimated proportion of users willing to use the ith technique. Espino et al. (2007) showed that while a sweep net was more costreliable than visual inspection for sampling stink bugs in rice, low acceptance ( = 0.4, or 40%) by rice scouts rendered visual inspection superior.

7.3.6 Sample selection procedures The method used to select sample units has important implications for convenience, cost, reliability and precision. Cochran (1977) distinguished between probability and non-probability selection methods. A probability method is one where the chance that an individual unit is included in a sample is known, whereas probability of inclusion is not known with non-probability methods. Probability methods are more desirable because resulting estimates of means, variances and SEs (Table 7.1) can be assumed to be unbiased in most cases. In IPM sampling, it is convenient for scouts to collect a haphazard sample of units, or a systematic one by walking a habitat in an “V,” “W” or “X” pattern in an effort to get a “representative” sample. While these may be practical routines, they allow many opportunities for bias. Of particular concern are “edge effects,” where the target organism may be concentrated in field margins or centers. Also, unrecognized spatial patterns in density may coincide with the walked path. These non-probability methods can seriously under- or overestimate mean field abundance (see Alexander et al., 2005). Wherever practical, enhanced reliability of probability methods justifies their use. Simple or unrestricted random sampling This design (e.g. Fig. 7.4A) is the simplest and most widely used of the probability designs. Each sample unit has an equal chance (1/N) of being observed. The change over haphazard selection is that a randomization method is used to choose units. A variety of mechanical or digital methods can be used to select units. Points in twodimensional habitats can be located by randomly


Fig. 7.4 Illustrations of four sample selection procedures applied to a hypothetical 50 × 100 m field, subdivided into sample units of 1-m2 quadrats; N = 5000, n = 30. (A) Simple (unrestricted) random sample chosen with uniform random numbers for X–Y coordinates. Note parts of field undersampled. (B) Stratified sample with universe divided into edge (Ne = 1080, ne = 7) and center (Nc = 3920, nc = 23) to control for edge effect. (C) Two-stage sample of n = 15 primary units and m = 2 secondary units to reduce travel costs. (D) Systematic sample of units chosen in grid pattern, every 9th vertically, 12th or 13th horizontally.

choosing X–Y coordinates, mapping them (e.g. Fig. 7.4A), and then navigating through the spatially ordered list. If estimation or decision sampling are being done sequentially, then units can be mapped and processed in overlapping subgroups, each of which will require a separate pass through the habitat.

Stratified random sampling This design (e.g. Fig. 7.4B) is a powerful alternative to simple random sampling, because it assures coverage, makes use of the sampler’s knowledge, and often gives greater control over precision and cost. The universe is divided into an arbitrary number of parts (strata, h = 1 . . . L) that are mutually exclusive subgroups of units of known or estimated number, Nh . Strata can be horizontal, vertical or both; they can be of different sizes and

irregular in shape, and units in a stratum do not need to be contiguous. The goal is to form groups of units with means as different as possible, and whose internal variances are as small as possible. Unless the sampler is interested in comparing the strata themselves, division of a universe into more than six strata generally produces marginal improvements in overall precision and efficiency (Cochran, 1977). Once the universe is divided, each stratum is sampled using simple random sampling. Means and variances from each are calculated (as in Table 7.1) and then combined to obtain a weighted mean for the whole universe: x¯st =


wh x¯h = w1 x¯1 + w2 x¯2 + · · · w L x¯L ,




 L  s2 wh2 h (1 − fh ) . SEst =

nh h=1


The weights, wh = Nh /N, are the proportions of total N that are in the different strata. Effective number of degrees of freedom (df) for SEst will depend on sampling fractions and stratum variances, and will be between the smallest (nh – 1) and (n – 1) (Cochran, 1977). To use a stratified design, a sampler allocates a total n among the strata. The simplest




arrangement is known as proportional allocation, where the number of units per stratum is proportional to stratum size. Proportional allocation is a good choice if preliminary estimates of stratum variances are not available, and cost of processing a unit is approximately the same everywhere. Many field monitoring protocols instruct scouts to divide fields into quarters, and to inspect multiple points in each one. This is actually a stratified design, intended to assure field coverage, with proportional allocation of equal sample sizes (nh ) in four equally sized strata. However, this prescribed design may not be the most efficient one if units in the quarters have different variances, and if time to travel to the different quarters varies widely. A refinement over proportional allocation is known as optimal allocation, which adjusts for differences in stratum variances and sampling costs, and yields the greatest precision per unit sampling cost. Presuming preliminary estimates of variances and costs are available, sample size for individual strata are calculated from nˆh = n ×

√ wh s h c h , L  √ (wh s h c h )



where sh is square root of variance (standard deviation) in stratum h, and ch is cost of obtaining and processing a unit in stratum h. Inspection of Eq. (7.8) indicates that the optimal allocation calls for more units in a stratum if it is larger than the others, if it is more variable internally, or if its units are cheaper to obtain. In circumstances where a stratum’s variance is exceptionally great, the optimal allocation may call for every unit in that stratum to be examined, i.e. set nh = Nh . For example, stratified sampling and optimal allocation is used by the State of Minnesota, Department of Transportation (MN–DOT) to estimate abundance of Canada thistle (Cirsium arvense) (Moon, 2007a). Cluster sampling and multi-stage subsampling These designs were originally developed to survey humans in populations too large to be enumerated and sampled at random (see Lohr, 1999). Rather, blocks of households could be selected randomly from maps, and surveyors could travel to chosen blocks to interview all residents in each

block. This approach is now known as cluster sampling, where a simple random sample of blocks (= primary sample unit, 1◦ ) is chosen, and then each resident (secondary sample unit, 2◦ ) within the chosen blocks is surveyed. A variant of cluster sampling is known as multi-stage subsampling, where chosen clusters are randomly subsampled rather than examined in their entirety. Nesting with subsampling at each level is easily extended to three or more levels, as might occur with trees (1◦ s) in a plantation, branches (2◦ s) within trees, twigs (3◦ s) in branches and ultimate sample units of leaves (4◦ s) on twigs. For brevity, we will illustrate a two-stage design with primary units of equal size, as might occur if a field were divided into blocks of equal size (e.g. Fig. 7.4C). Readers should consult Cochran (1977) and Lohr (1999) for guidance if primary or lower units are unequal in size, or if three or more stages are required. Estimators for a two-stage nested design are complicated by the fact that sampling and variation occur at both 1◦ and 2◦ levels. To develop more notation, let xi,j represent an observation from i = 1 . . . n ≤ N 1◦ units, and j = 1 . . . m ≤ M possible 2◦ s per 1◦ . The overall sample mean is x¯

m n 1  xi, j , nm i=1 j=1


with SE =

(1 − f1 )

n s 12 s2 + (1 − f2 ) 2 . n N nm (7.10b)

Subscripts 1 or 2 for sampling fractions and sample variances refer to 1◦ or 2◦ units, respectively. The two variance components arise from variation among 1◦ units and among 2◦ units within 1◦ s; both components are estimated with a nested analysis of variance (ANOVA). Benefits of cluster sampling and subsampling in IPM are that the sampler can control sampling costs by reducing travel time between the n 1◦ units, collecting m 2◦ s in each one, for a total of nm observations in all. A shortcoming, though, is that precision with two-stage sampling is often less than if the same nm units were chosen independently with simple random sampling. This occurs because units in the same cluster tend


to be positively correlated, and consequently the extra effort to process more than a small subsample adds little information. Depending on the spatial scale of variability in the habitat, precision in multi-stage designs is most frequently determined by variability among 1◦ units. Guidance on the optimum number, mopt of 2◦ units per 1◦ unit that will yield the greatest precision per total cost can be obtained from 


mopt =  s 12 −

s 22

c1 ≈ c2

c1 s 22 × . s 12 c2


mean can be calculated as in Table 7.1. However, in situations where there is an underlying gradient or trend in density, then a systematic sample can yield a biased estimate of the mean. Worse, if there is a cyclical or periodic pattern that coincides with the interval t, then potential for bias is great. A further limitation of systematic samples is that, because not every sample of size n has an equal chance of being observed, sample variance can be biased, and resulting confidence intervals can be unreliable.


Here, c1 is cost to locate the average 1◦ unit, and c2 is cost to process the average 2◦ unit. Total cost would be C = nc1 + nmc2 . The optimal number of 2◦ units per 1◦ depends on the ratios of the two variance components and the two cost components; the more variable the 2◦ s and the cheaper they are to process, the larger mopt will be. Modifications of Eqs. (7.10 and 7.11) have been developed to incorporate different variance–mean relations among 1◦ and 2◦ levels, and variable c2 , which can also depend on mean density (Hutchison, 1994; Binns et al., 2000). In rare cases, calculated mopt can exceed M, and will occur if variance among 2◦ units exceeds M times the variance among 1◦ s. In this case, set mopt = M; variability among 2◦ units is so great, full cluster sampling will be more efficient than subsampling. Systematic designs These designs are convenient alternatives to simple random sampling, because they greatly simplify the task of physically locating sample units. They are also useful when the objective is to map pest density (Fleischer et al., 1999). If the N units can be arranged in a linear order, and the sampler desires a sample of size n, then the interval t between chosen units will be the next integer after N/n. To choose a specific random start sample, one chooses a random integer R within 1 . . . t, and then actually observes units numbered R, R + t, R + 2t, . . . , R + (n – 1)t (Lohr, 1999). An equivalent procedure in two-dimensional universes leads to grid patterns (e.g. Fig. 7.4D). If pest density is effectively independent of order, then a systematic sample will behave much like a simple random sample, and the sample

Combinations of the four basic designs Simple random, stratified, multi-stage and systematic designs can be used in combination. For example, a field could be stratified to achieve coverage, and then the strata could be sampled using a twostage nested design, with systematic selection of 2◦ units within randomly chosen 1◦ units. As long as numbers of units in each subdivision of a universe are known, then results can be weighted appropriately to obtain unbiased estimates. Interested readers should consult Cochran (1977) or Lohr (1999) for guidance.

7.4 Decision sampling This third form of sampling is fundamentally different from detection and estimation. Rather than detect a pest or estimate its abundance, the goal is to decide if abundance at a given time is safely below a critical threshold, or if it is above and justifies remedial action. For simplicity, we will refer to this decision as “act” versus “no-act,” and action may involve a variety of remedial responses. Thresholds can be based on expert opinion or deeper knowledge (see Chapter 3). The utility of decision sampling is greatest in managed habitats where pests are intermittently below or above threshold, where losses can be great and where remedial actions are costly or potentially disruptive to the habitat or to crop market value. Assuming the threshold is correct, decision sampling will lead to a correct decision with a predetermined error rate (e.g. 5%), and will do so with a minimum sample size (Binns et al., 2000). Decision sampling has been shown to reduce sampling costs by 40% to




60%, compared with fixed sample size (RSS) methods with equal type I error rate. Decision sampling was first developed by Abraham Wald to determine if the quality of munitions manufactured during World War II met acceptable standards. At the time, testers thought they could accept or reject a product well before the requisite number of tests had been completed, especially when failures were very rare or frequent. In response, Wald developed sequential probability ratio tests (SPRTs) that formalized the sequential decision process. His procedures were later adapted to make spray or no-spray decisions in forest pest management in the 1950s, and in row crops in the 1960s (see Binns, 1994). More recently, sequential decision methods have been extended to make three-choice decisions, to allow more flexibility in characterization of biological sampling distributions, and to use computers rather than analytical methods to evaluate performance and practicality of proposed plans. One approach is with Monte Carlo methods that use random number generators for specified probability distributions to simulate performance of proposed sequential plans (see Binns et al., 2000). A second approach is to use resampling methods to evaluate and adjust sampling plans, based on repeated resampling of data sets obtained from extensive field samples (Naranjo & Hutchison, 1997).

7.4.1 General process Decision sampling requires a sample unit that is biologically appropriate and convenient, a defined threshold, and rules for terminating sampling and making a decision with specified error rates. We will designate a threshold as a critical proportion of units infested, pc , or as number of individuals per sample unit, µc , depending on whether organisms in units are scored for presence/absence or are fully counted. In practice, the sampler begins by examining a minimum number of units, nmin , scores each,  calculates the total, Tn = xi , and then updates Tn as additional units are examined. Sampling stops and a decision is made when Tn first exceeds the range between lower or upper “stop limits,” as listed in a table or graph. When abundance is well below or above the threshold, a decision can be

made early, because Tn quickly exceeds the corresponding limit. In contrast, when true abundance is near the threshold, a decision may not be possible without a very large sample. To prevent sample size from becoming prohibitively large, most sequential plans set an upper limit on sample size, nmax . If reached before a decision is made, then the recommendation is to take no action, but return soon to resample.

7.4.2 Stop limits Wald framed a decision problem as two competing one-tailed tests, Ho: µ < µ1 and Ha: µ ≥ µ2 , where (µ1 + µ2 )/2 = µc . If Ho is true, no action will be taken, whereas if Ha is true, then action would be warranted. Unfortunately, chance sampling makes it possible that two kinds of errors can occur. An estimated mean (= Tn /n) can be above µ2 when the true mean is below µ1 ; the consequence would be to waste action when not warranted. The converse error is that an estimate can appear to be below µ1 when truly above µ2 . In this case, the decision would be to wrongly do nothing, and incur a loss that could have been prevented. These errors are known as a type I and type II errors, respectively, and convention is to designate their probabilities as α and β. Wald derived stop line equations for binomial, normal, Poisson and negative binomial distributions, and these are tabulated in Binns et al. (2000). For all four distributions, the stop lines are increasing functions of n. When graphed in relation to n (Fig. 7.5A), they appear as parallel lines straddling n × pc or n × µc and are separated vertically by an amount that depends on chosen levels for α and β. A conceptual limitation of Wald’s approach is that his µ1 and µ2 are arbitrary lower and upper critical levels, whereas critical limits in IPM are singular. As an alternative to Wald’s SPRTs, Iwao (1975) proposed that upper and lower limits could be derived from confidence intervals around the threshold. For pc , Iwao’s upper and lower decision lines are    U n ≈ n pc + zα+2 s 2 /n


 L n ≈ n( pc − zα/2 s 2 /n),




which also are centered around the diagonal line, n×pc , but diverge slightly (Fig. 7.5A). The standard normal two-tailed deviate for chosen α does not reliably control type I error rate, but it does govern distance between the two boundaries. Iwao’s method does not explicitly control type II error rate. Estimated variance, s2 is for p at pc . Other sampling distributions can be accommodated by substituting appropriate estimators for s2 (see Table 7.1).

7.4.3 Performance of decision sampling plans

Fig. 7.5 Properties of two sequential decision plans for spider mites on cotton, with pc = 0.80. Dotted lines: Wald’s binomial SPRT, p0 = 0.75, p1 = 0.85, α = β = 0.05. Solid lines: Iwao’s procedure pc = 0.80, α = 0.05. Limits of nmin set at 5 leaves and nmax at 75. (A) Calculated stop limits, with results of two simulations with hypothetical p(infested) = 0.9 or 0.6. (B) Operating characteristic (OC) curves. (C) Average sample number (ASN) curves. OC and ASN curves estimated with simulations of 1000 trials at hypothetical infestation levels between 0.1 to 1.0.

A workable plan must lead to a decision over a range of true pest densities surrounding the threshold. This aspect of performance is displayed by an operating characteristic (OC) curve (Fig. 7.5B), which shows the probability of taking no action over a range in abundance straddling the threshold; the steeper the curve at the threshold, the better the plan. A workable plan must also be affordable over the anticipated range in pest density. This aspect is revealed in a plot of average sample number (ASN) versus hypothetical density (Fig. 7.5C), which is the expected number of sample units that will be required to reach a decision at a given density. To maintain chosen error rates ( = reliability), expected sample size must increase as abundance nears the threshold. Wald provided equations for calculating OC and ASN curves as functions of density for his four SPRTs, whereas curves for Iwao’s procedure must be derived through computer simulation (see Binns et al., 2000). Important elements of a specific decision sampling plan are its error rates. How do you set an acceptable level for α, the probability of erroneously taking action, and β, the probability of erroneously not taking action? Wilson (1982) offered guidance. For α, the appropriate error rate occurs when the combined cost of sampling, remedial action, and ecological disruption are at a minimum. For β, the appropriate error rate occurs when the combined cost of sampling, action, disruption and loss due to pest injury are at a minimum. To illustrate, three species of spider mites are commonly found in cotton in the USA, and their damage potential depends on when




populations become established in a given field. Spider mites can complete as many as 16 generations per season, depending on season length. A practical sample unit for spider mites is an individual cotton leaf, one per randomly chosen plant. Leaves are scored as clean or infested, because mites are far too numerous to be counted individually. The economic injury level for spider mites varies with species and crop phenology, but a workable threshold is pc = 0.80 ± 0.05 leaves infested (Wilson & Room, 1983). Here, we set Type I and II error rates at 0.05, and nmin and nmax at 5 and 75 leaves, respectively. Sequential plans for spider mites using these parameters and Wald’s and Iwao’s procedures are summarized in Fig. 7.5. In practice, a scout enters a field at weekly intervals during the period of highest risk and observes a leaf from each of five plants, and totals the number infested. If the total is outside the stop limits, then sampling stops and a decision is made to either do nothing or to apply an acaricide, depending on whether the lower or upper limit was exceeded. On the other hand, if the total is within the central “indecision zone,” then additional leaves are inspected sequentially until a decision can be made. To limit the possibility that sampling might continue indefinitely, sampling is halted after a predetermined 75 leaves have been examined. If the total at that point is still in the indecision zone, it is recommended that the field be resampled in three days. For comparison, RSS = 64 leaves (via Eq. 7.6a) would be needed to estimate p when 0.8 with h = 0.05, α = 0.05. In this example, Wald’s stop limits are slightly broader than Iwao’s when sample sizes are below 25, but narrower when above (Fig. 7.5A). The OC curves indicate the two plans would have equivalent “no-act” profiles (Fig. 7.5B). Average sample number (ASN) curves with the two plans are equivalent at low infestation levels, but ASNs with Iwao’s plan are greater than with Wald’s plan when infestation rate is near or above threshold (Fig. 7.5C). Despite minor differences in performance, either plan would enable scouts to make treatment decisions with far fewer leaf inspections than would be required by a fixed-precision plan with equal type I error rate.

7.5 Practical considerations and future needs The elements of IPM sampling discussed thus far should offer guidance in the design and use of sampling plans for individual pests on a one-byone basis. In practice, though, most IPM situations involve more than one potential pest species. Furthermore, the pest fauna and flora is likely to change as the growing season progresses within any region, and from one geographic region to another. A practical difficulty for IPM scouts is that the optimal sample unit and sampling design for one species is not likely to be the same for all. Hence, compromises must be made if multiple pests are to be monitored simultaneously. Solutions are currently being found by practitioners through common sense and intuition. There is a need to develop an economic framework for optimizing sampling plans for multiple species situations, based on sampling costs and economic value of the information obtained. IPM researchers and stakeholders working in some agricultural systems have invested to develop sampling programs that are practical and effective. However, Wearing (1988) surveyed IPM researchers and educators, and found that about 50 percent of respondents in Australia, New Zealand, Europe and the USA ranked “lack of simple monitoring methods” as the greatest technical barrier to IPM implementation. We are unaware of formal surveys that have assessed sampling practices in any IPM sector, but we suspect rigor and extent of adoption are greatest where the economic and environmental stakes are highest, but much room for improvement remains. Unfortunately, sampling requires labor that is costly in many parts of the world, so there is a universal need to find technologies that could reduce labor costs and improve efficiency. There is growing interest in remote sensing, forecasting models and global positioning systems to direct sampling resources to locations and times of greatest need. Additionally, web-based data entry, analysis and reporting procedures may ease adoption of rigorous sampling procedures, and permit retrospective analyses of their performance.


References Alexander, C. J., Holland, J. H., Winder, L., Woolley, C. & Perry, J. (2005). Performance of sampling strategies in the presence of known spatial patterns. Annals of Applied Biology, 146, 361–370. Binns, M. R. (1994). Sequential sampling for classifying pest status. In Handbook of Sampling Methods for Arthropods in Agriculture, eds. L. P. Pedigo & G. D. Buntin, pp. 137–174. Boca Raton, FL: CRC Press. Binns, M. R., Nyrop, J. P. & van der Werf, W. (2000). Sampling and Monitoring in Crop Protection: The Theoretical Basis for Developing Practical Decision Guides. Wallingford, UK: CABI Publishing. Cochran, W. G. (1977). Sampling Techniques. New York: John Wiley. Espino, L. A, Way, M. O. & Wilson, L. T. (2007). Determination of Oebalus pugnax (Hemiptera: Pentatomidae) spatial pattern in rice and development of visual sampling methods and population sampling plans. Journal of Economic Entomology, 101, 216–225. Fleischer, S. J., Blom, P. E. & Weisz, R. (1999). Sampling in precision IPM: when the objective is a map. Phytopathology, 89, 1112–1118. Hutchison, W. D. (1994). Sequential sampling to determine population density. In Handbook of Sampling Methods for Arthropods in Agriculture, eds. L. P. Pedigo & G. D. Buntin, pp. 207–243. Boca Raton, FL: CRC Press. Iwao, S. (1975). A new method of sequential sampling to classify populations relative to a critical density. Research in Population Ecology, 16, 281–288. Karandinos, M. G. (1976). Optimum sample size and comments on some published formulae. Bulletin of the Entomological Society of America, 22, 417–421. Kogan, M. & Herzog, D. C. (eds.) (1980). Sampling Methods in Soybean Entomology. New York: Springer-Verlag. Kuno, E. (1991). Verifying zero-infestation in pest control: a simple sequential test based on the succession of zero-samples. Research in Population Ecology, 33, 29– 32. Lohr, S. L. (1999). Sampling: Design and Analysis. Pacific Grove, CA: Brooks/Cole Publishing. Madden, L. V. & Hughes, G. (1999). Sampling for plant disease incidence. Phytopathology, 89, 1088–1103. Moon, R. D. (2007a). Estimation of Canada thistle areas along roadway right-of-ways. In Radcliffe’s IPM World Textbook, eds. E. B. Radcliffe, W. D. Hutchison & R.

E. Cancelado. St. Paul, MN: University of Minnesota. Available at moon1.htm. Moon, R. D. (2007b). Detection of leafy spurge along roadway right-of-ways. In Radcliffe’s IPM World Textbook, eds. E. B. Radcliffe, W. D. Hutchison & R. E. Cancelado. St. Paul, MN: University of Minnesota. Available at Naranjo, S. E. & Hutchison, W. D. (1997). Validation of arthropod sampling plans using a resampling approach: software and analysis. American Entomologist, 43, 48–57. Pedigo, L. P. & Buntin, G. D. (eds.) (1994). Handbook of Sampling Methods for Arthropods in Agriculture. Boca Raton, FL: CRC Press. Perry, J. N. (1981). Taylor’s power law for dependence of variance on mean in animal populations. Applied Statistics, 30, 254–263. Ruesink, W. G. (1980). Introduction to sampling theory. In Sampling Methods in Soybean Entomology, eds. M. Kogan & D. C. Herzog, pp. 61–78. New York: SpringerVerlag. Southwood, T. R. E. & Henderson, P. A. (2000). Ecological Methods, 3rd edn. London: Blackwell Science. Taylor, L. R. (1961). Aggregation, variance and the mean. Nature, 189, 732–735. Taylor, L. R., Woiwod, I. P. & Perry, J. N. (1978). The density-dependence of spatial behavior and the rarity of randomness. Journal of Animal Ecology, 47, 383–406. Venette, R. C., Moon, R. D. & Hutchison, W. D. (2002). Strategies and statistics of sampling for rare individuals. Annual Review of Entomology, 47, 143–174. Wearing, C. H. (1988). Evaluating the IPM implementation process. Annual Review of Entomology, 33, 17–38. Wilson, L. T. (1982). Development of an optimal monitoring program in cotton: emphasis on spider mites and Heliothis spp. Entomophaga, 27, 45–50. Wilson, L. T. (1994). Estimating abundance, impact, and interactions among arthropods in cotton agroecosystems. In Handbook of Sampling Methods for Arthropods in Agriculture, eds. L. P. Pedigo & G. D. Buntin, pp. 475– 514. Boca Raton, FL: CRC Press. Wilson, L. T. & Room, P. M. (1983). Clumping patterns of fruit and arthropods in cotton, with implications for binomial sampling. Environmental Entomology, 12, 50–54.


Chapter 8

Application of aerobiology to IPM Scott A. Isard, David A. Mortensen, Shelby J. Fleischer and Erick D. De Wolf Insects, plant pathogens and weeds that move through the air create some of the most interesting pest management problems because their populations can increase dramatically, often with little warning and independent of factors that operate within fields (Irwin, 1999; Jeger, 1999). The advent of IPM programs has created an increased need to predict when, where and which pest populations are likely to grow rapidly and require control. Where dispersal is critical to the dynamics of populations, the realization of this demand requires information on the movements of pests into and out of agricultural fields and the degree to which fields within landscapes and regions are interconnected by these flows (Isard & Gage, 2001). Fundamental to this need is a solid understanding of aerobiology, the study of the biological and atmospheric factors that interact to govern aerial movements of biota (aerobiota) among geographic places (Aylor & Irwin, 1999). Aerobiology, and dispersal research in general, is currently “on the move,” in large part due to rapid advances in technologies for measuring and analyzing flows of organisms at relevant temporal and spatial scales (Gage et al., 1999; Westbrook & Isard, 1999; Blackburn, 2006).1 The renewed


attention to issues of movement spans a wide range of pest and beneficial taxa that use air, water and/or land to change position on Earth for a multiplicity of reasons. Human-mediated dispersal is receiving much attention as well, although inadvertent pest movements in association with the globalization of commodity exchanges and human travel are extremely difficult to measure (National Research Council, 2002). The success of IPM tactics may be limited by insufficient knowledge of aerobiology (Irwin & Isard, 1994). Being able to estimate the contribution of immigrants to a pest population, versus those derived from within-field development, and whether these immigrants are derived locally or from more distant sources, would dramatically improve our ability to manage pest density and resistance. Programs to introduce biological enemies would also benefit from improved predictions of aerial movements of these organisms. Technologies to bio-engineer crops and other organisms are advancing rapidly, but little is understood about how these engineered organisms respond to natural fluctuations of weather and their ability to spread through the air. This knowledge is necessary to meet

See special sections in Agricultural and Forest Meteorology, 1999, vol. 97, no. 4; Ecology, 2003, vol. 84, no. 8; Diversity and Distributions, 2005, vol. 11, no. 1; and Science, 2006, vol. 313, no. 5788.

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


Fig. 8.1 The original version of the conceptual model was called the aerobiology pathway model and the stages were source, release, dispersion, deposition and impact. Figure adapted from Benninghoff & Edmonds (1972). The stages were renamed by Isard & Gage (2001) to extend the applicability of the conceptual model across a range of aerobiota from microorganisms to birds.

regulatory requirements for deploying genetically engineered organisms (Rissler & Mellon, 1996), and to design effective resistance management programs (Brown & Hovmoller, 2002). The potential risk from the airborne introductions of engineered animal and plant pathogens, pollen and seeds is also considerable (Wheelis et al., 2002; Stack et al., 2006). Whether these organisms are released purposefully or by mistake, the same aerobiological knowledge is required to understand and predict their spread.

8.1 The aerobiology process model and scales of movement During the US Biophysical Program on Aerobiology in the 1970s, the systems approach began to be employed to focus research on the many interrelated factors that influence the movement of aerobiota including agricultural pests (Edmonds, 1979). From this collaborative effort, a conceptual model of the aerobiology process emerged (Fig. 8.1) which still serves as a foundation for aerobiological research today (Isard & Gage, 2001). The model holds that each organism that changes geographic location by moving through the air proceeds sequentially through five basic stages: (1) preconditioning, (2) takeoff and ascent, (3) horizontal transport, (4) descent and landing and (5) impact.

Knowledge of the atmospheric medium through which organisms move is fundamental to understanding the suite of ecological and environmental factors that they experience in each airborne life stage and consequently the impact they cause at their destination (Pedgley, 1982; Aylor, 2003). Although the aerobiology process model is scale independent, at any given time and place, the air is composed of a variety of atmospheric motion systems that are hierarchically nested in time and space (Stull, 1988). For example, a typical hurricane/midlatitude cyclone (macroscale) is composed of many thunderstorms/tornadoes (mesoscale) and a multitude of gusts/eddies (microscale). Organisms may encounter a number of these motion systems in a single journey. Consequently, the temporal and spatial scales relevant for organisms that move in the air are highly variable ranging from seconds to weeks and from plant parts to continents (Isard & Gage, 2001). The biology and ecology of organisms that exploit atmospheric winds to disperse, within and among habitats, dictate the relevant temporal and spatial scales for particular individuals and populations and strongly influences their patterns of movement (Nathan, 2005). A taxonomy of movements useful for arthropods was constructed by Dingle (1996) based on the work of John S. Kennedy. Relating movement to life histories, Dingle grouped movement behaviors into three types: (1) those that are directed by a home range or are resource dependent, (2) those that are not directly responsive to resource or home range




and (3) those that are not under the control of the organism. Movements that meet the second criteria are considered to be migratory. Although this classification does not apply to plant pathogens and weeds that disperse in the air, it is extremely useful for characterizing commonalities in migratory behaviors among arthropod pests and for identifying the spatial and temporal scales associated with their movements. Phenological, physiological and morphological adaptations for movement in the air expressed by plant pathogens and weeds also can be used to identify relevant scales of movement. Examples include the plant controlled traits that dictate terminal velocity, release height and abscission in seeds (Tackenberg et al., 2003), liberation mechanisms in fungi (Gregory, 1973), and pigmentation in spores that influence survival to UV radiation exposure during aerial transport (Aylor, 2003). Price (1997, ch. 3) provides a synthesis of how the size and shape of organisms influence how they relate to physical forces and thus their movement behaviors, with examples that encompass organisms relevant for IPM applications. Development of these movementadapted traits and behaviors among individuals in populations in source regions (preconditioning) are crucial to the initiation of movement and the status of the biota during the aerobiological process. As a result of the structure of atmospheric motion systems and the biology of organisms that use the air to move, three scales of study are particularly relevant to the application of aerobiology in IPM: field, landscape and continental. At the field scale, the architecture of the plant canopy is shaped by interplant competition. The resulting canopy architecture strongly influences the temperature and moisture conditions as well as the small-scale motions within the atmospheric medium through which organisms move (Lowry & Lowry, 1989). Individual journeys occurring within a canopy or immediately above a field are typically shorter than 100 m and last a few minutes at most. An overwhelming proportion of passively transported organisms move only a few meters from their point of origin in a single transport event (Gregory, 1973). Because populations of pests that utilize crops for resources can build rapidly, and because growers typically only have

realistic opportunities to manage fields, the individual field has been the spatial scale at which IPM is most often practiced. While in many situations dispersal occurs at a field scale for much of a growing season, movements among landscape elements are often associated with behaviors related to colonization and overwintering. Management opportunities exist at these wider scales (National Research Council, 1996; Hunter, 2002) and in many circumstances aerobiology is a critical component of effective ecologically based management at scales larger than a single field. Pheromone disruption programs directed at insect pests, for example, should operate at scales that encompass the matesearching area (Jones, 1998). Areawide efforts at pest management, programs aimed at altering the rate of geographic expansion (Tobin et al., 2004) and eradication efforts directed against early-stage invasive pests require understanding of movement among host resources and especially between agricultural and surrounding landscapes (Ekbom, 2000). An understanding of movement biology is also key to developing effective strategies for using transgenic crops for areawide suppression of pest populations (Carriere et al., 2003) and for successful pro-active deployment of resistance management strategies for transgenic crops (Sisterson et al., 2005). Local winds that blow across landscapes are influenced by the physical characteristics of the Earth’s surface and the extent of the geographic assemblages of ecosystems that comprise the landscape (Oke, 1987). Relief in a landscape can induce mechanical winds while thermal circulations typically occur during calm atmospheric conditions when skies are clear and where there is spatial variation in the rate of surface heating. These landscape-scale winds provide opportunities for pests to move among diverse habitats (Nathan, 2005). Generally journeys of individuals span hundreds to thousands of meters and occur within a single diel period (Isard & Gage, 2001). To a large extent, landscape-scale movements of populations occur over relatively short time periods (hours to days), either because the life history of the species makes many individuals ready to move at the same time and/or because only a limited number of atmospheric motion systems


facilitate movement to appropriate destinations (Pedgley, 1982). The resultant flows of organisms both among fields with similar crops and among diverse habitats within landscapes are often critical to the dynamics of pest populations in agricultural fields and provide opportunities for novel IPM practices. At the continental scale, weather systems embedded in the global circulation govern the aerial movement of organisms (Pedgley et al., 1995). These systems range from thunderstorms to large cyclones that traverse tens to thousands of kilometers over days to weeks. The progression of spring from the subtropics into midlatitude continents (i.e. warm temperatures and longer days) triggers emergence and development of many plants and animals (Gage et al., 1999). Large-scale advection of warm moist air from the subtropics at this time provides a mechanism for long-distance transport of organisms that are in dispersal-ready life stages from low latitudes to more poleward regions (Johnson, 1995). In summer, when the latitudinal temperature gradient is less pronounced, winds are weaker and opportunities for continental-scale biological movements are less frequent. However, opportunities for long-distance movement increase again in early autumn with the advent of tropical cyclones and hurricanes (Isard et al., 2005). The incursion of cold polar air masses in autumn also stimulates the senescence of most crops and occurs at a time when many weedy species have mature seeds preconditioned for flight. At this time, enhanced latitudinal temperature gradients create weather systems with strong winds directed from the poles equatorward that are capable of transporting dispersal ready aerobiota from the middle latitudes to overwintering habitats in the subtropics (Johnson, 1995). Either because of their biology or due to the nesting of landscape and localscale motion systems within large-scale weather patterns, or both, other organism such as seeds exploit continental-scale weather patterns to disperse over lesser distances. Sudden and often dramatic influxes of pests associated with these largescale weather systems occur each spring in most midlatitude North American agricultural regions (Stakman & Harrar, 1957), and similar largescale movements toward subtropical latitudes are

facilitated by weather patterns during the fall (Showers et al., 1993).

8.2 Application of aerobiology principles to IPM in the twenty-first century Realizing the difficulties of making generalizations about the movement process because of the biological diversity of aerobiota and the relevance of events that occur at multiple and often nested temporal and spatial scales, a group of scientists and outreach specialists met in 1993 to create an aerobiology research focus (Isard, 1993; Isard & Gage, 2001). They developed a generic set of hypotheses pertaining to long-distance transport of biota during the middle three stages of the aerobiology process (Table 8.1). Although the hypotheses pertain to movement in the atmosphere, the participants acknowledged that longdistance movement can only be understood holistically if one understands what happens at each end of the movement process as well. Their vision was that researchers would evaluate the generic hypotheses across a wide range of both biological and meteorological systems. The resulting scientific principles would then provide the basis for applying aerobiology to IPM and ecosystem management in general. The technologies for understanding and using aerobiology to enhance IPM decision support have advanced dramatically. Enabling technologies supported by the cyber infrastructure that link weather data and remotely sensed phenomena to pest movement, and enhance coordination of on-the-ground sampling of pests and/or hosts and real-time exchange of expert opinion by practitioners have made it possible to functionally integrate risk assessment across multiple disciplines and data sources to address complex pest management issues. The technologies that have enhanced observation, analysis and communication capabilities include: (1) Measurement, monitoring and diagnostics tools that allow rapid and more accurate spatially referenced data collection, (2) use of citizen scientists for expanding our biological observations networks over large




Table 8.1

General purpose hypotheses governing the atmospheric movement of biota

Maintenance of the movement process (1) Long-distance movement is a one-way process (2) Long-distance movement is a two-way process (2a) Return movement is ancillary to long-distance aerial transport (2b) Return movement reinforces the genetic control over long-distance transport (2c) Return movement is driven by existing environmental conditions Components of the movement process (A) Takeoff and ascent (3) Takeoff and ascent into the atmosphere by organisms that move long distances is biologically mediated (3a) The phenological state that invokes initiation of ascent is genetically controlled (3b) Environmental preconditioning induces a physiological state that causes initiation of ascent (3c) Intraspecific/interspecific interactions influence the initiation of ascent (3d) Ascent may be influenced by aerodynamic properties (4) Ascent by organisms into the atmosphere is influenced by environmental conditions (4a) Ascent is governed by convection within the lower atmosphere (4b) Thresholds of important atmospheric factors limit the tendency to take off and the success of ascent (4c) Ascent may be caused by hydrometeors (B) Horizontal transport (5) Organisms are concentrated within atmospheric layers during long-distance aerial movement (5a) Behavior and aerodynamic properties govern the vertical distribution of organisms during long-distance aerial movement (5b) Atmospheric factors govern the vertical distribution of organisms during long-distance aerial movement (6) Horizontal transport of organisms within the atmosphere is predictable (6a) The duration and direction of movement are determined by the organism (6b) The duration and direction of movement are affected or driven by atmospheric processes (6c) The duration and direction of movement are influenced by environmental preconditioning (C) Descent and landing (7) Organisms actively descend from the atmosphere (7a) Environmental cues induce descent (7b) Physiological status govern descent (7c) Intraspecific/interspecific interactions influence the initiation of descent (8) The descent of organisms from the atmosphere is governed by meteorological factors (8a) Descent is caused by hydrometeors (8b) Descent is caused by downdrafts (8c) Descent is caused by changes in stability/turbulence (8d) Descent is caused by gravity Source: Taken from Isard (1993).

regions, (3) internet-based tools for data collection that enable immediate sharing of observations from around the world, (4) remote sensing instrumentation for monitoring important environmental variables (e.g. weather and land cover)

that allow us to run models in near real time and to forecast change, (5) high-speed computing that enhances our capacity to add value to observations through modeling, (6) spatial analysis tools for higher levels of synthesis of processes that occur


over space and time and (7) internet-based tools for rapid communication of commentary and guidelines for managing pest problems. In the case studies that follow, we demonstrate the usefulness of applying the aerobiology process model in IPM. We describe novel types of pest management problems that operate at landscape to continental scales, and how advanced technologies can improve IPM decision making in the twenty-first century. To highlight the diversity of applications of aerobiology to IPM, we have chosen to focus on a combination of insect, plant pathogen and weed systems. The three case studies also represent different stages in the application of aerobiology to IPM. In the first, we related how aerobiology-based sampling methods have been used to increase our understanding of seed movement across landscapes and improve the effectiveness of management strategies for herbicideresistant weeds in genetically modified crops. The second case study describes a network for monitoring migratory noctuid moths north of their overwintering range, which begins to provide early warnings and insights into the movement processes and seasonal patterns of this lepidopteran pest. In the final case study, we demonstrate the utility of coupling a system for forecasting the aerial spread of an important plant pathogen at the continental scale with an extensive monitoring system on a state-of-the-art information technology (IT) platform to provide real-time communications and IPM decision support.

8.3 Case studies 8.3.1 An aerobiology approach influencing herbicide resistance management plans Horseweed (Conyza canadensis, syn. Erigeron canadensis) is a winter annual plant, native to North America. Its distribution is worldwide, though it is more abundant in temperate climates. It is found in a wide range of soil types and disturbance regimes from coarse sandy soils to organic soils and agricultural fields to roadsides and recently abandoned fields (Leroux et al., 1996). Horseweed seedlings emerge from late August through October, forming rosettes that

overwinter (Holm et al., 1997), to emerge in late winter and early spring. The plant bolts in late spring, blooms in mid-July, and seed set occurs in early August. A mature plant can reach 2 m in height producing upwards of 200 000 seeds that are wind-borne with the aid of a pappus (Weaver, 2001). In the summer of 1998, horseweed resistant to glyphosate herbicide was found infesting soybean fields along the Delaware–Virginia coast (VanGessel, 2001). It was the first weed in an annual crop to develop resistance to glyphosate herbicide. This event was particularly important as glyphosate tolerance was introduced in genetically modified soybean and made commercially available in 1996. The adoption rate for this transgenic crop is unprecedented. In the 2002 field season approximately 75% of USA soybean hectares were planted to glyphosate-tolerant cultivars and the use rate has now reached 87% of planted hectares (Chassy, 2005; Dauer et al., 2006). Glyphosate-tolerant maize, alfalfa and cotton have since been commercialized and those crops have also experienced high adoption rates. Already, by 2005, 31% of the maize hectares in New York State had been planted to glyphosatetolerant maize (National Agricultural Statistics Service, 2007). Repeated use of a high mortality practice selects for resistance to that practice. The selection pressure for resistance to glyphosate is high given that it is regularly used in most of the widely planted summer grain and fiber crops. Complete reliance on glyphosate for weed control goes against the fundamental tenet of IPM to promote the use of a diversity of tactics to avoid selection for adapted species. Preventing the appearance of new glyphosate-resistant biotypes and limiting the spatial extent and impact of existing resistant populations is a pressing IPM challenge with profound agronomic and economic implications. While this discussion centers on resistant C. canadensis spread among fields, the underlying processes apply to many weedy species. In general, weed invasions are successful when the species can scatter seeds among heterogeneous environments, establish, adapt and then disperse again to colonize other areas (Sakai et al., 2001). Wind dispersal provides species with an effective method




of dispersing throughout a landscape. Some common agricultural weeds that are increasing on farmlands in the USA, including thistles (Carduus spp. and Cirsium spp.), dandelion (Taraxacum officinale), milkweed (Asclepias syriaca) and C. canadensis, use the atmosphere to move throughout landscapes and often long distances. Establishment in and adaptation to the varying cropping systems within which weed seeds land are critical next steps in the invasion process. Individuals that successfully establish themselves in new locations play an important role in the invasion because they are adapted to the destination habitat and act as seed sources for the next wave of the invasion. The increase in no-tillage hectares can be directly linked to adoption of glyphosate-tolerant crops and the result is a reduction in herbicide heterogeneity (Carpenter et al., 2002). This aids in the establishment of resistant horseweed and the invasion of additional farms. Consequently, resistant horseweed provides an ideal subject for the study of long-distance dispersal across complex landscapes. Aerobiological knowledge gained through this research effort in turn provides a basis for developing management strategies for many of our most problematic weeds. Seed dispersal is a dynamic process combining functions of plant and seed morphology, habitat characteristics (e.g. topography and rugosity), vector mediation and post-settlement dynamics (Cousens & Mortimer, 1995). Horseweed produces small seeds with an attached pappus. The terminal fall velocity (settling velocity) of a seed is dramatically reduced by the presence of a plume structure (Burrows, 1975). Anderson (1993) measured settling velocities of wind-dispersed seeds from 19 species of the family Asteraceae, and results for horseweed were the lowest of those tested. Combined, these characteristics indicate a high potential for wind dispersal of horseweed over considerable distances. To better understand just how far seeds could move, thus defining the interconnectedness of fields within an agricultural landscape, a series of experiments were initiated. Source patch populations were established and the source strength of each patch was quantified. Seed traps were positioned along radial transects from the source patches out to 500 m. The results of the study are striking (Dauer et al., 2006). Small num-

bers of seeds were found near the ends of transects revealing that it was likely that they were moving beyond 500 m. This is five times greater than any previously reported dispersal distance for horseweed. Simulations conducted using a twodimensional model and a range of realistic source strength values, indicated that it is highly likely that horseweed can disperse distances of 2–5 km. Not surprisingly, the dispersal distance was greatest in the direction of the prevailing wind direction. The results reveal that farm fields are much more highly interconnected with respect to weed management outcomes than previously thought. Prior to these aerobiology-based studies, herbicide resistance management and the development of weed infestations were considered field-specific processes. To confirm this novel and important finding, Shields et al. (2006) deployed remotely controlled aircraft to sample seeds in the air above and downwind of source patches. Specifically, they were interested in whether or not seeds are able to leave the surface boundary layer of the atmosphere and attain an altitude where they are likely to be blown long distances. In this campaign, horseweed seeds were collected at the highest altitude sampled (110 m above the field). Concomitant sampling of seeds and measurements of wind speed, wind direction and turbulence immediately above the source patch provide the basis for aerobiological analyses. The greatest numbers of horseweed seeds were captured above the surface boundary layer, generally considered to be twice the height of the canopy in the field (Oke, 1987), during the mid afternoon. It is reasonable to assume that seeds which are carried upward through the surface layer during midday by thermals are mixed throughout the planetary boundary layer above by evening when turbulence in the lower atmosphere typically diminishes and the air becomes stable. In this situation, given a fall velocity for horseweed of 0.32 m/s (Dauer et al., 2006), the seeds would require between 3 min and 1 h to settle out of the air. Assuming that the seeds were released near midday and landed 1 h after dusk, a light wind speed of 5 m/s (11 mph) would transport them between 75 and 150 km while a strong wind of 20 m/s (43 mph) would transport them over 550 km (Shields et al., 2006).


The Conyza system and management problem provided an opportunity to apply aerobiologybased methods for sampling a pest organism and associated environment to enlighten our understanding of how to manage herbicide-resistant weeds in genetically modified crops. The fact that seed dispersal distances may stretch many kilometers is important in a number of ways. First, it demonstrates that plants possess the ability to move their propagules and novel genes much greater distances than previously thought. Such movement has important implications for the biology and ecology of the species. From an IPM perspective, such long-distance transport effectively decouples a farmer’s management decision from the resulting weed control. A neighbor’s field that contains a wind-dispersed weed will continue to act as a source for invasion into adjacent fields regardless of the management practices used to maximize control in adjacent fields. In the case of glyphosate-resistant weeds, the continued invasion could limit the options of farmers utilizing glyphosate-tolerant crops. One solution is cooperation among farmers to slow the spread of this and other wind-dispersed species. Already such cooperation is evident at the level of the commercial applicator and consultant. In regions on the scale of commercial applicator districts, resistance management herbicide programs have been devised and deployed. Invariably, they involve either adding a pre-emergence treatment prior to the glyphosate application that targets horseweed or using an additional postemergence herbicide with a different mode of action to control the glyphosate-resistant horseweed. Clearly, applying the knowledge of interfield movement of horseweed seed to growing regions that have not witnessed the establishment of glyphosate-resistant horseweed populations would be the most prudent action in the long run. This is because managing resistant populations in ways mentioned above bring with it $US 20–37 per hectare increase in herbicide cost and the environmental impact of increased herbicide load. In a recent symposium addressing the subject of managing glyphosate-resistant weeds, Mortensen et al. (2007) argued that other areawide management practices should also be considered and could include establishing buffer regions

around fields or farmsteads with resistant biotypes. Within the buffer, a more diverse cropping practice and herbicide program would be employed. For example, recently VanGessel et al. (2007) reported that winter cover crops like rye and hairy vetch significantly limit the establishment and fecundity of glyphosate-resistant Conyza. We argue that detailed aerobiology research can be used to guide pesticide registration and agricultural policy as well as stewardship practices. For example, if spread of glyphosateresistant biotypes occurs over large distances and over short time intervals, such information could help the US Environmental Protection Agency make decisions about future registrations of crops carrying this trait. For example, the impact of increasing the number of hectares on which glyphosate can be used by decreasing management practice diversity was explored using a spatially explicit modeling approach and with knowledge of the horseweed dispersal direction and distance (Mortensen et al., 2003). The invasion speed of the resistant horseweed increased dramatically when it was assumed that the maize and alfalfa hectares in a Pennsylvania study region were planted to glyphosate-tolerant cultivars. Clearly herbicide management plans that target the individual field and grower will not contain the spread of problem weeds such as horseweed that move well beyond a farmstead in a single day.

8.3.2 An aerobiological approach influencing measurement of migratory lepidopterans Insect species unable, or less able, to adapt to the harsh winters that cover most of the North American continent are eliminated or dramatically reduced in all or most of the region. Examples include lepidopterans in the family Noctuidae, e.g. the corn earworm (Helicoverpa zea), fall armyworm (Spodoptera frugiperda), black cutworm (Agrotis ipsilon), beet armyworm (Spodoptera exigua) and soybean looper (Pseudoplusia includens). These migratory lepidopterans annually reinvade their northern geographic range through longdistance aerial movements involving successive broods advancing northward (Holland et al., 2006). The corn earworm, in the subfamily Heliothinidae, serves as a model where understanding its




population dynamics requires an aerobiology knowledge base and should be considered at a continental scale. The biology and ecology of corn earworm enable it to use both resource-directed and migratory dispersal to exploit dispersed host resources. Adaptations influencing this dispersal behavior can be placed in context of its evolutionary history. Prior to 1953, and for over 100 years, there were only one or two recognized taxa in this subfamily, all in Heliothis (asaulta, armigera, or obsoleta) with typically global distribution. A distinct New World taxon (zea) was recognized in the 1950s (Common, 1953; Todd, 1955), which Hardwick (1965) placed into a newly erected genus Helicoverpa. Heliocoverpa, considered the ancestral genus (Hardwick, 1965), retains much higher fecundity (hundreds to thousands of eggs) than the other major groups (Heliothis or Schinia), and a much wider host range. Derived groups have decreasing fecundity and increased specializations, such as the ability to oviposit within floral structures, which are tied to a more specific host range and perhaps more resource-directed dispersal. Helicoverpa, in contrast, widely disperses many eggs, and has retained a much wider host range, along with stronger migratory behaviors. This ancestral group is comprised primarily of tropical or warmtemperate species, with continuous development in tropical regions except where there is a dry season. Some members (zea and armigera) have limited diapause capabilities in higher latitudes; none do this well. Neither zea from the Americas nor armigera from the Old World overwinters above 40◦ latitude (Hardwick, 1965). Thus, H. zea annual reinvasions affect the entire North American continental interior extending into Canada. A similar process occurs with armigera in the Old World, and allozyme patterns suggest that H. zea represents a founding event from H. armigera (Mallet et al., 1993). Holland et al. (2006) suggest that insects are often facultative migrants, responding to existing or predicted changes in habitat quality to achieve “multi-generational bet-hedging.” “Bet-hedging” for H. zea could involve portions of the population exhibiting local, resource-directed redistribution among nearby hosts in the landscape, while others exhibit long-range migratory flight. Landscape-

scale redistribution is well documented and has been modeled as an ovipositional choice process weighted by distance among fields (Stinner et al., 1986). Long-range migratory flight is also documented (Lingren et al., 1994; Westbrook et al., 1997). Probability of recaptured moths carrying a citrus-pollen marker, indicating migration, was modeled as a logistic function of the insect’s flight trajectory, duration of flight the previous night, and local minimum air temperature (Westbrook et al., 1997). Aerobiology analyses suggest that migratory moths utilize dynamic but definable pathways, and synoptic weather patterns correlate with corn earworm migration into southern (Lingren et al., 1994; Westbrook et al., 1997) and midwestern USA (Sandstrom et al., 2005). As with other migratory noctuids (e.g. black cutworm: see Sappington & Showers, 1991), corn earworm displays behavior that distinguishes migration from foraging, including a spiraling vertical ascent behavior at dusk (Lingren et al., 1995) which carries them above the boundary layer to altitudes of >800 m where they often concentrate in nocturnal wind jets associated with temperature inversions (Wolf et al., 1990; Beerwinkle et al., 1995). Within these jets, moths often orient at an angle to the downwind direction, perhaps to influence migration direction (Wolf et al., 1995), compensate for wind drift, or reduce fallout. Airborne radar indicated that this persistent and straightened-out movement enables earworms to move a few hundred kilometers in a single night (Wolf et al., 1990). Such distinctive poleward migratory flights during spring have been documented (Lingren et al., 1994), while there is indirect evidence for return equatorward fall flights (Gould et al., 2002). Mark–release–recapture of black cutworm provide direct evidence for both poleward spring and equatorward fall movements (Showers et al., 1989, 1993). Air parcel trajectory analyses and radiosonde measurements indicated that these moths move long distances at altitudes between 300 and 800 m (Showers, 1997). Black cutworm also shows an age dependency for migratory flights (Sappington & Showers, 1991) with flight initiation inhibited by specific weather conditions (Domino et al., 1983). Thus, corn earworm is well adapted to our ephemeral cropping systems, and achieves pest



Fig. 8.2 Interactive map from PestWatch (Fleischer et al. 2007) showing captures of corn earworm for August 10 in the drought year of 2002 (A) compared to 2001 (B). Histogram insets show distribution of average daily capture for the current week among ∼250 sites. Maps depicting changes over time may be viewed as animations while time-series from individual dates can be viewed as still frames at each sampling site. Geographic extents were expanded in 2007.

status in maize, cotton, tomatoes, snap beans, sorghum and soybeans. Larvae feed on flowers, fruits and seeds, boring quickly into reproductive tissue, causing direct damage to economically and nutritionally important plant parts. Management with insecticides or biological control is difficult once larvae access the interior of fruiting bodies. Coupled with a tendency to distribute eggs singly among fruiting structures, damage can become economically significant very quickly and can be disproportionately high relative to corn earworm density.

North of their overwintering range, dramatic influxes of corn earworm and other lepidopteran migrants can result in rapid population increase independent of factors that operate within fields. The timing and intensity of invasive, migratory flights can vary dramatically. For example (Fig. 8.2), in 2001, regional densities primarily from migrants into the northeastern USA had just begun to increase by approximately 10 August, whereas in the next year, 2002, much larger populations were evident throughout the region at the same time. Consequently, regional monitoring networks, a component of aerobiological approaches to IPM, have direct relevance. These include enhanced observation, analysis and communication capabilities. They can be used to rapidly and effectively visualize pest dynamics over wide geographic regions, at varied temporal and spatial scales, enabling us to discuss the migratory process as a comprehensive whole, and thus consider migratory processes when we consider management options. Aerobiological





Fig. 8.2 (cont.)

structuring of data flows enables early warnings, both from historical knowledge gained through organized observations structured with information technologies, and from modeled forecasts that couple pest, host and meteorological processes. Autonomous monitoring of corn earworm has occurred for decades for IPM of cotton, sweet corn or other vegetable crops. In the northeastern USA, almost entirely north of the overwintering range of the corn earworm, the migratory contribution to the density and dynamics of this pest made it critical to consider pest movement to improve IPM. A regional monitoring network was established across seven northeastern states in 1999 to provide advance warning about this migratory process (Fleischer et al., 2007), and this network has recently expanded to approximately 500 sites. The maps influence insecticide spray decisions in vegetable crops.

8.3.3 An aerobiology based decision support system for managing soybean rust In autumn 2004, the Asian strain of soybean rust (Phakopsora pachyrhizi), an exceptionally virulent fungus, was found infesting soybean fields in Louisiana (Schneider et al., 2005). During the preceding years, the US Department of Agriculture (USDA) had prepared for the incursion of this pathogen by supporting grower, specialist and industry education, training, and surveillance programs, purchases of new equipment for diagnostic facilities, offshore fungicide evaluation trials, construction of risk assessments and searching breeding materials for novel sources of host plant resistance (Livingston et al., 2004). The sense of urgency stemmed not only because P. pachyrhizi had demonstrated frequent long-distance aerial spread among all other major soybean growing areas worldwide, but also because yield losses


from infected fields had been significant in each of these production regions (Miles et al., 2003). In 2003, shortly after soybean rust had blown from Africa across the South Atlantic Ocean to infect soybean fields in Brazil, Paraguay and Argentina, a number of research groups began developing models to assess the risk of P. pachyrhizi incursion into the USA, likely transport pathways, and when and where deposition of spores was most likely to occur (Isard et al., 2005). At the onset of this program, the aerobiology process model (Fig. 8.1) was specified for the soybean rust system to identify relationships that needed to be incorporated into the transport models as well as to identify knowledge gaps that were used to guide field measurement programs in South America and the USA for evaluating aerobiology-related hypotheses (e.g. Isard et al., 2006a). Output from the resulting Integrated Aerobiology Modeling System (IAMS) became the basis for the USDA, Economic Research Service assessment of the risk of a soybean rust incursion and its potential impact on USA agriculture (Livingston et al., 2004). IAMS simulations conducted immediately after discovery of soybean rust in the USA showed that airflows converging into Hurricane Ivan as it made landfall along the Gulf Coast had the potential to transport viable rust spores directly to the USA from the infected Rio Cauca source area in Colombia, South America. A model generated map delineating regions of spore deposition associated with the hurricane was provided to members of the USDA Animal and Plant Health Inspection Service (APHIS) Soybean Rust Rapid Response Team as they went into the field and was used successfully to scout for the pathogen (see Fig. 5 in Isard et al., 2005). Phakopsora pachyrhizi is an obligate parasite that requires green tissue to survive (Bromfield, 1984). In North America, it has primarily been found on two hosts, soybean and kudzu. Consequently, the geographic range of soybean rust during winter is restricted to areas along the Gulf Coast, and in the Caribbean basin, Mexico and Central America where either kudzu retains its foliage or soybeans are grown year round (Pivonia & Yang, 2004). To cause yield losses in major North American soybean production regions, P. pachyrhizi uridineospores must be blown from these overwinter-

ing areas into the continental interior between early May and August, when the crop is susceptible to the disease (Isard et al., 2005). During the 2004/2005 winter, APHIS provided funds to the IAMS team for the construction and operation of an IT platform to integrate soybean rust monitoring, database construction and communications to stakeholders (Isard et al., 2006b). The team used the opportunity to integrate emerging technologies into a unique and highly functional, state-of-the-art cyber infrastructure. The level of cooperation among USDA agencies, state Departments of Agriculture, universities, industry and grower organizations, in support of the resulting USDA Soybean Rust Information System was unprecedented for an invasive agricultural pest in the USA, enabling the deployment of a pest information system with an exceptional level of utility and credibility. As a result of this success, government administrators, researchers, industry representatives and producers employed the same template the following year to launch a national Pest Information Platform for Extension and Education (PIPE) including soybean aphid along with soybean rust (Isard et al., 2006b). The IAMS is configured in a modular format to include all of the stages in the aerobiology process (Isard et al., 2007). Host development and disease progression submodels, driven by weather data, are used to characterize source strength and distribution as well as colonization and disease progression at destinations. The IAMS also incorporates information on spore release and canopy escape in source areas, mortality due to exposure to solar radiation during atmospheric transport and both dry and wet deposition. During the 2005 and 2006 growing seasons, the IAMS modeling team provided maps of output from daily simulations depicting deposition of spores on soybean and kudzu, host developmental stage and disease progression on these hosts for the subtropical and temperate regions of North America. Observations from an extensive network of “sentinel” soybean plots were used to define the geographic extent and severity of the disease in source areas. Because of the complexity of the biological and environmental interactions that are important to the soybean rust system and the paucity of knowledge about this




Fig. 8.3 The information technology structure of the Pest Information Platform for Extension and Education (PIPE) as designed for soybean rust. The observation component in the lower portion of the figure pertains to a plant pathogen and is different for insects and weeds.

disease in North America, one major limitation of the IAMS initially was the need to “nudge” the model. That is, to produce realistic simulations of pathogen transport and subsequent disease detection, model parameter values had to be continually adjusted and field observations of vegetative growth and changes in soybean rust distribution had to be continually fed into the IAMS. To achieve model adjustment in near real time and to provide a platform upon which to communicate interpretations of field observations and model output to users, the PIPE cyber infrastructure was developed. The PIPE is people connected by computers using advanced IT to add value to field observations. The goals of the system are to enhance support for IPM decision making and provide information for documenting good management practices for crop insurance purposes. The IT structure of the PIPE is depicted in Fig. 8.3. Field observations of the pathogen, hosts and environment are channeled through standardized internet portals into a national database. This was accomplished by the developing protocols for field monitoring of the pathogen and its hosts and building a set of tools including PDA programs, internet forms and spreadsheet files for easy and rapid entry of observations into the national database.

National Weather Service products are also downloaded daily and archived. These spatiotemporalreferenced data are immediately available to researchers throughout the country who are able to add value to the observations through modeling. Outputs from multiple models running on the platform are integrated with field observations of soybean rust into easy to read maps. Participants access information on the PIPE through two interrelated websites. A restricted access website provides a platform for extension specialists, researchers and administrators to view and interpret these maps. Extension specialists then use state-of-the-art IT tools to disseminate interpretations, management guidelines and other relevant materials to growers and industry agents through a public access website (Isard et al., 2006b). Realizing the success of the PIPE for providing useful information to support decision making for soybean rust IPM, the USDA Risk Management Agency (RMA), the Cooperative State Research Extension and Education Service (CSREES) and APHIS are developing a coordinated effort to maintain and expand the system capitalizing on the existing structure of regional IPM Centers and state extension specialists (Isard et al., 2006b). The IPM Centers set national and regional pest priorities through interactions with industry and as a result of their leadership, the PIPE information technology structure was expanded in 2006 to include soybean aphid. Plans for future additions to the platform include viral diseases of dry beans, head scab of barley and lepidopteran pests of sweet corn. Even more critical to the success of the PIPE is the network of state extension specialists who coordinate the input of observations from monitoring networks into the national database and interpret field observations and model output to provide agricultural producers and industry with decision support for pest management and information for documenting their management activities.

8.4 Conclusions Successful IPM implementation requires conceptual models that integrate knowledge of the biology, ecology and behavior of host–pest complexes


(Jeger, 1999). Where aerial dispersal of pests is an important component of the system, knowledge of the scale at which movements occur, and the flow structures within the atmospheric environment within which they take place, enable us to most effectively utilize ecologically based pest management. Understanding movement is absolutely essential for management that restricts geographic distribution of pests, alters genetic frequencies in populations, and thus potentially causes pesticide resistance. The Conyza case study demonstrates how aerobiology-based research can fill knowledge gaps and thus enhance understanding of movement across landscapes; this in turn, can improve the effectiveness of our management strategies for herbicide-resistant weeds in genetically modified crops. In the corn earworm case study, IT tools are applied to spatially referenced monitoring of a pest for which migratory processes cause rapid increases in population independent of factors that operate within fields. The development of early warning systems based on regional monitoring networks such as PestWatch (Fleischer et al., 2007), are a prerequisite for running and verifying aerobiology models that forecast impending movements. There are important feedback loops throughout this process. Efforts to visualize, describe and model movement of pests also reveal knowledge gaps and aerobiological methods can help direct future research priorities. Also, the information about pest density, dynamics and phenology at wider scales adds value to data collected at individual sites, thus increasing the effectiveness and efficiencies of scouting efforts. The soybean rust example shows how an extensive biological observation network, coupled with mechanistic aerobiology transport models and state-of-the-art communications capabilities, all operating over large geographic scales can serve IPM. In 2006, the USDA Economic Research Service reported that many millions of USA soybean hectares that would otherwise have received unnecessary fungicide application for soybean rust in 2005 remained untreated for soybean rust in 2005 due to this application of aerobiology for IPM of soybean rust. In that year alone, the information disseminated through the USDA Soybean Rust Information System increased USA produc-

ers’ profits between $US 11 to 299 million ($US 0.40 to 10.18 per hectare) at a cost that was only a fraction of this return (Roberts et al., 2006).

References Anderson, M. C. (1993) Diaspore morphology and seed dispersal in several wind-dispersed Asteraceae. American Journal of Botany, 80, 487–492. Aylor, D. E. (2003). Spread of plant disease on a continental scale: role of aerial dispersal of pathogens. Ecology, 84, 1989–1997. Aylor, D. E. & Irwin, M. E. (1999). Aerial dispersal of pests and pathogens: implications for integrated pest management. Agricultural and Forest Meteorology, 97, 233– 234. Beerwinkle, K. R., Lopez, J. D. Jr., Schleider, P. G. & Lingren, P. D. (1995). Annual patterns of aerial insect densities at altitudes from 500 to 2400 meters in east-central Texas indicated by continuouslyoperating vertically-oriented radar. Southwestern Entomology (Suppl.), 10, 63–79. Benninghoff, W. S. & Edmonds, R. L. (eds.) (1972). Ecological Systems Approaches to Aerobiology. I. Identification of Component Elements and Their Functional Relationships, United States/International Biophysical Program Aerobiology Program Handbook No. 2. Ann Arbor, MI: University of Michigan. Blackburn, L. (2006). Tag team. Science, 313, 780–781. Bromfield, K. R., 1984. Soybean Rust, Monograph No. 11. St. Paul, MN: American Phytopathological Society Press. Brown, J. K. M. & Hovmoller, M. S. (2002). Aerial dispersal of pathogens on the global and continental scales and its impact on plant disease. Science, 297, 537–542. Burrows, F. M., 1975. Calculation of the primary trajectories of dust seeds, spores and pollen in unsteady winds. New Phytologist, 75, 389–403. Carpenter, J., Felsot, A., Goode, T. et al. (2002). Comparative Environmental Impacts of Biotechnology-Derived and Traditional Soybean, Corn and Cotton Crops. Ames, IA: Council for Agricultural Science and Technology. Carriere, Y., Ellers-Kirk, C., Sisterson, M. S. et al. (2003). Long-term regional suppression of pink bollworm by Bacillus thuringiensis cotton. Proceedings of the National Academy of Sciences of the USA, 100, 1519–1523. Chassy, B. (2005). Crop Biotechnology and the Future of Food: A Scientific Assessment. Ames, IA: Council for Agricultural Science and Technology. Common, I. F. B. (1953). The Australian species of Heliothis (Lepidoptera, Noctuidae) and their pest status. Australian Journal of Zoology, 1, 319.




Cousens, R. & Mortimer, A. M. (1995). Dynamics of Weed Populations. Cambridge, UK: Cambridge University Press. Dauer, J. T., Mortensen, D. A. & VanGessel, M. (2006). Spatial and temporal dynamics governing long distance dispersal of Conyza canadensis. Journal of Applied Ecology, 44, 105–114. Dingle, H. (1996). Migration: The Biology of Life on the Move. Oxford, UK: Oxford University Press. Domino, R. P., Showers, W. B., Taylor, E. & Shaw, R. H. (1983). Spring weather pattern associated with suspected black cutworm moth (Lepidoptera: Noctuidae) introduction into Iowa. Environmental Entomology, 12, 1863–1871. Ekbom, B. (2000). Interchanges of insects between agricultural and surrounding landscapes. In Interchanges of Insects between Agricultural and Surrounding Landscapes, eds. B. Ekbom, M. E. Irwin & Y. Robert, pp. 1–3. Dordrecht, Netherlands: Kluwer. Edmonds, R. L. (ed.) (1979). Aerobiology: The Ecological Systems Approach. Stroudsburg, PA: Dowden, Hutchinson & Ross. Fleischer, S. Payne, G., Kuhar, T. et al. (2007). Pestwatch sweetcorn pest monitoring system. In Pestwatch, eds. S. Fleischer et al. University Park, PA: Pennsylvania State University. Available at Gage, S. H., Isard, S. A. & Colunga-Garcia, M. (1999). Biological scales of motion for dispersal of biota. Agricultural and Forest Meteorology, 97, 249–261. Gould, F., Blair, N., Reid, M. et al. (2002). Bacillus thuringiensis-toxin resistance management: stable carbon isotope assessment of alternate host use by Heliocoverpa zea. Proceedings of the National Academy of Science of the USA, 99, 16 581–16 586. Gregory, P. H (1973). The Microbiology of the Atmosphere. New York: John Wiley. Hardwick, D. F. (1965). The Corn Earworm Complex, Memoir No. 40. Ottawa, Canada: Entomological Society of Canada. Holland, R. A., Wikelski, M. & Wilcove, D. S. (2006). How and why do insects migrate? Science, 313, 794–796. Holm, L., Doll, J., Holm, E., Pancho, J. & Herberger, J. (1997). Conyza canadensis (L.) Cronq. (syn Erigeron canadensis L.). In World Weeds: Natural Histories and Distribution, pp. 226–235. New York: John Wiley. Hunter, M. (2002). Landscape structure, habitat fragmentation, and the ecology of insects. Agricultural and Forest Entomology, 4, 159–166. Irwin, M. E. (1999). Implications of movement in developing and deploying integrated pest management strategies. Agricultural and Forest Meteorology, 97, 235– 248.

Irwin, M. E. & Isard, S. A. (1994). Movement and dispersal and IPM. In Proceedings of the 2nd National Integrated Pest Management Symposium/Workshop, pp. 136– 137. Raleigh, NC: Experiment Station Committee on Organization and Policy, Pest Management Strategies Subcommittee and Extension Service IPM Task Force. Isard, S. A. (ed.) (1993). Alliance for Aerobiology Research Workshop Report. Champaign, IL: Alliance for Aerobiology Research Workshop Writing Committee. Isard, S. A. & Gage, S. H. (2001). Flow of Life in the Atmosphere: An Airscape Approach to Understanding Invasive Organisms. East Lansing, MI: Michigan State University Press. Isard, S. A., Gage, S. H., Comtois, P. & Russo, J. M. (2005). Principles of aerobiology applied to soybean rust as an invasive species. BioScience, 55, 851–861. Isard, S. A., Dufault, N. S., Miles, M. R. et al. (2006a). The effect of solar irradiance on the mortality of Phakopsora pachyrhizi urediniospores. Plant Disease, 90, 941– 945. Isard, S. A., Russo, J. M. & De Wolf, E. D. (2006b). The Establishment of a National Pest Information Platform for Extension and Education. Online, Plant Health Progress, doi:10. 1094/PHP-2006–0915-01-RV, available at Isard, S. A., Russo, J. M. & Ariatti, A. (2007). Aerial transport of soybean rust spores into the Ohio River Valley during September 2006. Aerobiologia, 23, 271–281. DOI 10.1007/s10453-007-9073-z. Jeger, M. J. (1999). Improved understanding of dispersal in crop pest and disease management: current status and future directions. Agricultural and Forest Meteorology, 97, 331–349. Johnson, S. J. (1995). Insect migration in North America: synoptic-scale transport in a highly seasonal environment. In Insect Migration: Tracking Resources through Space and Time, eds. V. A. Drake & A. G. Gatehouse, pp. 31–66. Cambridge, UK: Cambridge University Press. Jones, O. T. (1998). Practical applications of pheromones and other semiochemicals. In Insect Pheromones and Their Use in Pest Management, eds. P. E. Howse, I. D. R. Stevens & O. T. Jones, pp. 263–351. London: Chapman & Hall. Leroux, G. D., Benoit, D. L. & Banville, S. (1996). Effect of crop rotations on weed control, Bidens cernua and Erigeron canadensis populations, and carrot yields in organic soils. Crop Protection, 15, 171–178. Lingren, P. D., Westbrook, J. K., Bryant, V. M. Jr. et al. (1994). Origin of corn earworm (Lepidoptera: Noctuidae) migrants as determined by Citrus pollen markers


and synoptic weather systems. Environmental Entomology, 23, 562–570. Lingren, P. D., Raulston, J. R., Popham, T. W. et al. (1995). Flight behavior of corn earworm (Lepidoptera: Noctuidae) moths under low wind speed conditions. Environmental Entomology, 24, 851–860. Livingston, M., Johansson, R., Daberkow, S. et al. (2004). Economic and Policy Implications of Wind-Borne Entry of Asian Soybean Rust into the United States. Electronic Outlook Report from the US Department of Agriculture, Economic Research Service, OCS-04D-02. Available at OCS04D02.pdf. Lowry, W. P. & Lowry, P. P. II. (1989). Fundamentals of Biometeorology: Interactions of Organisms and the Atmosphere, vol. 1. McMinnville, OR: Peavine. Mallet, J., Korman, A., Heckel, D. G. & King, P. (1993). Biochemical genetics of Heliothis and Helicoverpa (Lepidoptera: Noctuidae) and evidence for a founder event in Helicoverpa zea. Annals of the Entomological Society of America, 86, 189–197. Miles, M. R., Frederick, R. D. & Hartman, G. L. (2003). Soybean Rust: Is the US Soybean Crop at Risk? St. Paul, MN: American Phytopathological Society. APS Feature Story 06/2003, available at online/feature/rust/. Mortensen, D. A., Humston, R., Jones, B. & Dauer, J. (2003). A landscape analysis of an invasive composite. Weed Science Society of America, 43, 24 (abstract). Mortensen, D. A., Dauer, J. T. & Curran, W. S. (2007). Do ecological insights inform Conyza management? Northeastern Weed Science Society Abstracts, 61, 117. Nathan, R. (2005). Long-distance dispersal research: building a network of yellow brick roads. Diversity and Distributions, 11, 125–130. National Agricultural Statistics Service (2007). Agricultural Chemical Use Database. Washington, DC: US Department of Agriculture. Available at www. National Research Council (1996). Ecologically Based Pest Management: New Solutions for a New Century. Washington, DC: National Academy Press. National Research Council (2002). Predicting Invasions of Nonindigenous Plant and Plant Pests. Washington, DC: National Academy Press. Oke, T. R. (1987). Boundary Layer Climates. London: Routledge. Pedgley, D. E. (1982). Windborne Pests and Diseases: Meteorology of Airborne Organisms. Chichester, UK: Ellis Horwood. Pedgley, D. E., Reynolds, D. R. & Tatchell, G. M. (1995). Long-range insect migration in relation to climate and

weather: Africa and Europe. In Insect Migration: Tracking Resources through Space and Time, eds. V. A. Drake & G. A. Gatehouse, pp. 3–29. Cambridge, UK: Cambridge University Press. Pivonia, S. & Yang, X. B. (2004). Assessment of the potential year-round establishment of soybean rust throughout the world. Plant Disease, 88, 523–529. Price, P. (1997). Insect Ecology, 3rd edn. New York: John Wiley. Rissler, J. & Mellon, M. (1996). The Ecological Risks of Engineered Crops. Cambridge, MA: MIT Press. Roberts, M. J., Schimmelpfennig, D., Ashley, E. & Livingston, M. (2006). The Value of Plant Disease EarlyWarning Systems: A Case Study of USDA’s Soybean Rust Coordinated Framework, Economic Research Report No. 18. Washington, DC: US Department of Agriculture Economic Research Service. Available at Sakai, A. K., Allendorf, F. W., Holt, J. S. et al. (2001). The population biology of invasive species. Annual Review of Ecology and Systematics, 32, 305–332. Sandstrom, M., Chagnon, D. & Flood, B. R. (2005). How weather and climate impact your pest management decisions. In Vegetable Insect Management, eds. R. Foster & B. Flood, pp. 23–29. Willoughby, OH: Meister. Sappington, T. W. & Showers, W. B. (1991). Implications for migration of age-related variation in flight behavior of Agrotis ipsilon (Lepidoptera: Noctuidae). Annals of the Entomological Society of America, 84, 360–365. Schneider, R. W., Hollier, C. A., Whitam, H. K. et al. (2005). First report of soybean rust caused by Phakopsora pachyrhizi in the continental United States. Plant Disease, 89, 774. Shields, E. J., Dauer, J. T., VanGessel, M. J. & Neumann, G. (2006). Horseweed (Conyza canadensis) seed collected in the planetary boundary layer. Weed Science, 54, 1063– 1067. Showers, W. B., 1997. Migratory ecology of the black cutworm. Annual Review of Entomology, 42, 393–425. Showers, W. B., Whitford, F., Smelser, R. B. et al. (1989). Direct evidence for meteorologically driven longrange dispersal of an economically important moth. Ecology, 70, 987–992. Showers, W. B., Keaster, A. J., Raulston, J. R. et al. (1993). Mechanism of southward migration of a noctuid moth (Agrotis ipsilon Hufnagel): a complete migrant. Ecology, 74, 2303–2314. Sisterson, M. S., Carriere, Y., Dennehy, T. J. & Tabashniik, B. E. (2005). Evolution of resistance to transgenic crops: interactions between insect movement and field distribution. Journal of Economic Entomology, 98, 1751–1762.




Stack, J., Cardwell, K., Hammerschmidt, R. et al. (2006). The National Plant Diagnostic Network. Plant Disease, 90, 128–136. Stakman, E. C. & Harrar, J. C. (1957). Principles of Plant Pathology. New York: Ronald Press. Stinner, R. E., Saks, M. & Dohse, L. (1986). Modeling of agricultural pest displacement. In Insect Flight: Dispersal and Migration, ed. W. Danthariarayana, pp. 235– 241. Berlin, Germany: Springer-Verlag. Stull, R. B. (1988). An Introduction to Boundary Layer Meteorology. Dordrecht, Netherlands: Kluwer. Tackenberg, O., Poschlod, P. & Bonn, S. (2003). Assessment of wind dispersal potential in plant species. Ecological Monographs, 73, 191–205. Tobin, P. C., Sharov, A. A., Liebhold, A. A. et al. (2004). Management of the gypsy moth through a decision algorithm under the STS project. American Entomologist, 50, 200–209. Todd, E. L. (1955). The distribution and nomenclature of the corn earworm (Lepidoptera, Phalaenidae). Journal of Economic Entomology, 48, 600–603. VanGessel, M. J. (2001). Glyphosate-resistant horseweed from Delaware. Weed Science, 49, 703–705. VanGessel, M. J., Scott, B. A., Johnson, Q. R. & White, S. E. (2007). Horseweed response to no-till production

systems. Northeastern Weed Science Society Abstracts, 61, 116. Weaver, S. E. (2001). The biology of Canadian weeds. 115. Conyza canadensis. Canadian Journal of Plant Science, 81, 867–875. Westbrook, J. K., Wolf, W. W., Lingren, P. D. et al. (1997). Early-season migratory flights of corn earworm (Lepidoptera: Noctuidae). Environmental Entomology, 26, 12– 20. Westbrook, J. K. & Isard, S. A. (1999). Atmospheric scales of motion for dispersal of biota. Agricultural and Forest Meteorology, 97, 263–274. Wheelis, M., Casagrande, R. & Madden, L. V. (2002). Biological attack on agriculture: low-tech, high-impact bioterrorism. BioScience, 52, 569–576. Wolf, W. W., Westbrook, J. K., Raulston, J., Pair, S. D. & Hobbs, S. E. (1990). Recent airborne radar observations of migrant pests in the United States. Philosophical Transactions of the Royal Society of London B, 328, 619–630. Wolf, W. W., Westbrook, J. K., Pair, S. D., Raulston, J. R. & Lingren, P. D. (1995). Radar observations of orientation of noctuids migrating from corn fields in the Lower Rio Grande Valley. Southwestern Entomology (Suppl.), 18, 45–61.

Chapter 9

Introduction and augmentation of biological control agents Robert J. O’Neil and John J. Obrycki Natural enemies play key roles in pest management programs worldwide. Using natural enemies in pest management requires an understanding of their basic biology, how they impact pest population growth and how the environment and management system affect natural enemy dynamics and performance. Whereas there are abundant examples of using natural enemies in weed and plant pathogenic systems (VanDriesche & Bellows, 1996; Coombs et al., 2004), in this chapter we focus on insect pests and their associated predatory, parasitic and pathogenic natural enemies. Although natural enemies that occur naturally in crops can provide substantial control and are the cornerstone of many pest management programs, we will focus on approaches that intentionally add natural enemies to affect pest control. Examples of manipulating the environment to increase the number or effectiveness of natural enemies, sometimes referred to as conservation biological control, are given in several chapters in the current volume or in VanDriesche & Bellows (1996), Barbosa (1998) and Bellows & Fisher (1999). In this chapter we review approaches to introduce or augment natural enemies, provide case histories to illustrate their use and suggest research needs to increase the use of biological control agents in pest management. General texts on biological control (e.g. Debach & Rosen, 1991; VanDriesche & Bellows, 1996; Bellows & Fisher, 1999) provide a

broader overview of the methods and use of biological control in pest management systems.

9.1 Introducing new natural enemies The history of natural enemy introductions to control introduced pests, sometimes referred to as “classical biological control,” traces its origin to the successful use of vedalia beetle (Rodolia cardinalis), a predator coccinellid, to control the cottony cushion scale (Icerya purchasi) in California, USA in the 1880s (Debach & Rosen, 1991). That project identified many of the key tenets of classical biological control and was successfully replicated in most areas where cottony cushion scale had become a pest. The central idea behind the approach was that as an introduced insect (in this case from Australia), cottony cushion scale lacked the natural enemies in its new range that kept its population densities below economic levels in its homeland. Reuniting cottony cushion scale with its natural enemies would re-establish the natural enemy–prey dynamic, lower the pest population density and affect economic control. The steps to reunite so-called “exotic” natural enemies with an introduced pest include exploration in the pest’s area of origin,

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



Table 9.1

Examples of successful introductions of natural enemies as part of a classical biological control program

Pest (year)

Program costs (× $US 1000)

Sugarcane borer (1967) (Diatraea spp.) Coffee mealybug (1939) (Planococcus kenyae) Winter moth (1954) (Operophthera brumata) Spotted alfalfa aphid (1958) (Therioaphis maculata) Cottony cushion scale (1966) (Icerya purchasi) Olive parlatoria scale (1951) (Parlatoria oleae)

Benefit : cost (1st 10 years)













Source: After Gutierrez et al. (1999). quarantine examination for host specificity, release of the natural enemy into the environment and evaluation of the ecological and economic impacts of the introduction. Debach & Rosen (1991) estimated that classical biological control has resulted in the complete control of over 75 insect pests worldwide and substantial control of another 74 species. Since that publication additional successes have occurred in a variety of crop–pest systems. Accrued cost savings of successful projects can exceed many $US millions and obviate the need for further control interventions (Table 9.1). The environmental and health protections of classical biological control are less enumerated, but with often-significant reduction of insecticide use following successful biological control, they can be considerable. Although classical biological control has had a long track record of success, a number of studies have reported deleterious effects on non-target populations, including increased parasitism and predation rates and associated declines (and reported extinctions) of non-target host populations (Follett & Duan, 2000). Concern over possible negative effects of releases has resulted in more stringent evaluations of potential non-target impacts (see below). As with any pest control technology, the introduction of natural enemies is not risk-free, and evaluation of the risks associated with introducing natural enemies need to balance

the costs and benefits (both economic and ecological) of using natural enemies. Delfosse (2004) provides a framework to evaluate the relative risks of biological control (or other control technologies). Whereas earlier insect biological control programs paid scant attention to non-target impacts, current efforts include significantly more focus in this regard. (Non-target effects have been a concern for weed biological control for many years now.) We are encouraged that the history of biological control is replete with practitioners addressing challenges and modifying their approaches based on science and environmental concerns. We have every expectation that current and future practitioners will continue along these lines and expand the use of natural enemies in ever safer and more economic ways. Below we use several case histories to describe the use of natural enemy introduction to control insect pests. We have selected these projects to illustrate that natural enemies can be successfully used in a variety of crop–pest systems, and that ultimately success depends on commitment predicated on basic knowledge of pest and natural enemy ecology and systematics.

9.1.1 Cassava mealybug In the early 1970s farmers in Africa began noticing significant declines in cassava yields. Cassava, a crop introduced to Africa by Portuguese


colonists in the 1500s, is originally from the Americas. Grown as a subsistence crop, cassava provides essential nutrition to people in the “cassava belt” – an area south of the Sahel, and approximately the size of the continental USA. The cause of the crop loss was identified as a mealybug new to Africa, the cassava mealybug (Phenacoccus manihoti). Beginning in 1977, scientists at the International Institute of Tropical Agriculture (IITA) began a classical biological control program. Initially, explorations focused on Central America and northern South America as this was identified as the center of diversity of Phenacoccus. However, parasitoids reared from mealybugs collected in this area failed to attack cassava mealybug. Subsequent taxonomic revision suggested cassava mealybug was not from this area, and that other locations needed to be surveyed. Collections were then made in more southern latitudes in South America, where cassava mealybug was subsequently located. In 1981, a parasitoid, Apoanagyrus (= Epidinocaris) lopezi, was collected from cassava mealybug in Paraguay, and following host testing in quarantine (in Europe), releases were made in Nigeria. The parasitoid quickly established, spread throughout most of the cassava belt, and caused a significant decline in cassava mealybug densities and associated damage (Neuenschwander & Herren, 1988). It has been estimated that the cassava mealybug biological control project saved subsistence farmers in Africa hundreds of millions of dollars and helped secure food reserves for over 200 million people. The cassava mealybug project illustrates a number of points concerning classical biological control. First, the project shows how a fundamental understanding of the pest and natural enemy systematics and ecology is critical to success in importation biological control. While initial explorations focused on the area thought to be the most promising (because it was the center of diversity of the pest group), testing of natural enemies suggested additional areas should be explored. Exploration in the proper area led to the discovery of the key natural enemy and success in the project. The project also illustrates how using classical biological control is not limited by locale or the cropping system. Classical biological control had not been frequently attempted in

subsistence crops and it was argued that resourcepoor farmers in Africa would not benefit from the initial (c. $US 35 million) investment. Fortunately project personnel persevered and focused research to use biological control options for this introduced pest. The lead scientist on this project, Hans Herren, received the World Food Prize in 1995 for his leadership on the cassava management program. Finally, the success of the cassava mealybug led to work on other cassava pests, resulting in the biological control of the introduced cassava green mite (Mononychellus tanajoa) and later successes in other crops such as mangos and bananas.

9.1.2 Ash whitefly Ash whitefly (Siphoninus phillyreae) was introduced into California in the late 1980s. A pest of trees, including ash and ornamental pear, ash whitefly infestations in urban centers resulted in early season defoliation of street trees and subsequent loss of aesthetic value, and the cooling effects provided by shade. At high-density whitefly populations, respiratory health risks were reported, and outdoor activities declined due to “clouds” of flying adults. A biological control program was started by the California Department of Food and Agriculture, and the University of California (Pickett & Pitcairn, 1999). Within three years, a parasitoid, Encarsia inaron, was imported from Italy and Israel and released throughout most of the affected areas of California. Immediate and significant declines of ash whitefly populations were noted, and the whitefly is no longer considered a pest in California. Economic analysis of the program suggests $US 200–300 million savings in tree replacement costs and associated aesthetic benefits. The ash whitefly program illustrates the use of biological control in urban settings as well as in protecting the aesthetic value of the urban landscape. The project benefited by having an experienced set of researchers familiar with biological control, who quickly developed a biological control option. The project also benefited from past taxonomic work on the parasitoid taxon, as Encarsia species have been the focus of considerable work in biological control, both in classical and augmentative approaches (see below).




9.1.3 Soybean aphid Soybean aphid (Aphis glycines) was discovered infesting soybean fields in the Midwest USA in the summer of 2000. Having subsequently spread to 22 USA states and three Canadian provinces, the aphid directly reduces yield through feeding and is capable of transmitting a number of viruses. In outbreak years the pest has caused $US millions in damage, and an estimated 3 to 4 million hectares of soybeans have been treated with insecticides. Prior to the introduction of the soybean aphid, soybeans, particularly in the Midwest USA, were rarely treated with insecticides. The damage and associated control costs have made soybean aphid one of the most damaging insect pests of soybean production in the USA. In response to the threat posed by the soybean aphid a group of entomologists from several midwestern land-grant universities and research institutions, and the US Department of Agriculture initiated a classical biological control program. Foreign exploration for soybean aphid natural enemies began in 2001 and to date, nearly 30 populations representing six to nine species of parasitoids from several areas of northeast China, Korea and Japan have been received and successfully established in quarantines in the USA. Host specificity has been initially evaluated for at least 21 Asian populations. One species, Binodoxys communis, a newly described braconid, has undergone regulatory review and was approved for release in 2007. At the start of the program additional studies were initiated to assess potential nontarget impacts as part of a risk assessment of the biological control program. Included were incountry (Asia) field studies of host specificity of soybean aphid natural enemies, review of nontarget impacts of previous aphid biological control projects, development of a baseline assessment of aphid diversity using suction trap collections and field sampling of selected native aphid species and their associated natural enemies. Finally, a region-wide sampling program has been initiated to determine the impact of released natural enemies, both on the soybean aphid as well as other, non-target aphid species. The success of the soybean aphid project cannot be assessed at this time. However, it is important to note that the project began almost immedi-

ately following determination that the aphid was an introduced pest. Also, non-target impacts were of immediate concern, and studies were initiated and are an integral part of the research (Heimpel et al., 2004). Unfortunately, for all too many insect pests, classical biological control is not attempted, with the rationalization that the crop–pest system is not amenable to success. This is particularly true for annual crop systems that are thought to be too ephemeral for classical biological control to “work.” The soybean aphid project directly tests this perspective, and if successful will provide an additional case study to demonstrate that biological control in annual crop systems is a viable and important control option.

9.2 Augmentation Compared to introduction, in which the goal is the permanent establishment of a natural enemy to reduce the density of a pest, augmentative biological control involves periodic releases to reduce pest densities; permanent establishment of the natural enemy is not expected. As a method of biological control, augmentative tactics can be organized within two overlapping categories: environmental manipulations and periodic releases. The goal of augmentation, increasing the effectiveness of a natural enemy, is achieved by periodic releases of natural enemies or by manipulating the environment to favor the natural enemy. Over the past 40 years, overviews of the methods and results of augmentative releases have been published numerous times (e.g. Ridgway & Vinson, 1977; Parrella et al., 1999; van Lenteren & Bueno, 2003). In this chapter, we consider augmentation in a narrow perspective to include only periodic releases, and we highlight examples of successful augmentation programs, provide key references, outline methods and suggest areas for improvement. Readers interested in environmental manipulations (conservation biological control) can refer to several chapters in this book as well as VanDriesche & Bellows (1996), Barbosa (1998) and Bellows & Fisher (1999). Augmentative methods can involve government sponsored areawide management programs, grower-based cooperatives or individuals


making small-scale releases in gardens, greenhouses or urban landscapes. Regardless of the scale, augmentation biological control relies on the use of effective natural enemies, proper evaluation of efficacy and often integration with other control tactics (Ridgway & Vinson, 1977). Augmentative biological control has been used with semiochemicals for suppression of lepidopteran pests of cotton and tomatoes, chemical control for suppression of phytophagous mite pests in orchards and vineyards and plant resistance for management of whitefly pests of greenhousegrown cucumbers. Methods for release and evaluation of natural enemies are still a work in progress for many systems, although substantial progress has been made in the refinement of release protocols and economic evaluation of efficacy for selected environments such as greenhouse pest management (van Lenteren, 2000), and releases of Trichogramma wasps for control of lepidopteran pests in several systems (Smith, 1996). Augmentative releases can be inundative, when the natural enemy is released in high numbers to cause relatively rapid and direct mortality with no expectation of longer-term pest suppression. Examples of inundative programs include the use of entomopathogenic nematodes for suppression of the black vine weevil (Otiorhynchus sulcatus) in citrus, releases of high densities of Trichogramma targeting eggs of stalk boring lepidopteran pests of corn and sugarcane, and releases of immature Chrysopidae for reduction of homopteran pests (VanDriesche & Bellows, 1996; Barbosa, 1998). For many decades use of the bacterium Bacillus thuringiensis (Bt) was the most widely used pathogen for microbial control of lepidopteran pests based on an inundative strategy. Debates in the literature focused on questions of whether the use of Bt was a form of augmentative biological control or a biopesticide toxin delivery system. More recently, Bt genes have been incorporated into plants via genetic transformations, producing transgenic Bt corn, potato and cotton possessing a novel type of plant resistance (Shelton et al., 2002). The second type of augmentative release is inoculative, in which fewer numbers of natural enemies are released, with the expectation of longer-term (seasonal) pest suppression resulting

from the offspring of the released individuals. The ultimate goal of these periodic releases is to increase densities of the natural enemy, based on the underlying assumption that greater numbers of natural enemies result in greater suppression of the pest species. For example, the release of the parasitic wasp Encarsia formosa for suppression of the greenhouse whitefly (Trialeurodes vaporariorum) is based on increasing the densities of this parasitoid over the growing period of greenhouse floriculture and vegetable production systems. Pupae of the predatory fly Aphidoletes aphidimyza are released for aphid suppression which is accomplished via the production of larvae which consume the target prey. Similarly, reproduction by the mealybug destroyer (Cryptolaemus montrouzieri) is required for the production of predatory larvae which contribute to sustained biological control of mealybugs in controlled environments. Improvements in augmentation methodology include genetic selection of more efficient biotypes, pre-release conditioning of parasitoids through exposure to host materials, selection of appropriate natural enemy species for low or high pest densities and providing nourishment for released natural enemies. Several species of predatory mites have been selected for resistance against several pesticides commonly used for IPM in orchard systems. Studies demonstrated that individuals from releases of these resistant strains persisted in the environment when the pesticides were used in these systems. Pesticide-resistant strains of the western predatory mite (Galendromus occidentalis) are commercially available for biological control of spider mites in orchards and vineyards. Improved release methods can protect natural enemies from environmental mortality factors. For example, Trichogramma wasps may be enclosed in small, waxed cardboard capsules (TrichocapsR ) for efficient handling and to exclude predators. Each capsule contains approximately 500 parasitized Mediterranean flour moth (Ephestia kuehniella) eggs. Dormant immature Trichogramma within the Trichocaps can be stored at low temperatures until needed for field releases. Largescale field releases, timed to the seasonal phenology of the target pest, can be made due to the mechanized production and long-term storage of




Trichocaps. This type of mass production and storage provides hundreds of thousands of Trichocaps for field releases against the European corn borer (Ostrinia nubilalis) in western Europe (Obrycki et al., 1997). Commercial production of natural enemies is the major source of natural enemies used in augmentation in Europe and North America. For augmentative releases to be a viable component of pest management systems, natural enemies must be consistently mass produced and available. During the past two decades, the number of commercially produced natural enemy species has increased to over 125 species. An online list of North American suppliers of beneficial organisms for augmentative biological control has been compiled by Hunter (1997). General quality control guidelines for several species have been developed by the International Organization for Biological Control (van Lenteren et al., 2003). The biological parameters used to assess quality include the number alive in a container, adult size, longevity and fecundity, rates of emergence for natural enemies shipped as pupae, and estimates of rates of parasitism or predation. Quality control of commercially produced arthropod natural enemies remains a research focus within augmentative biological control (O’Neil et al., 1998).

9.2.1 Augmentative biological control in greenhouses Augmentation biological control is the foundation for many IPM systems for arthropod pests within the greenhouse environment (Parrella et al., 1999; van Lenteren, 2000). Two cosmopolitan plant pests of a wide variety of crops produced in greenhouses, two-spotted spider mite (Tetranychus urticae) and greenhouse whitefly are effectively managed through augmentative releases of the predatory mite (Phytoseiulus persimilis) and E. formosa, respectively. In the Netherlands, over 90% of greenhouse-grown tomatoes, cucumbers and peppers are produced using IPM systems based on augmentation (see Chapter 27). These successful programs in greenhouses evolved from a long-term commitment and sustained interactions between researchers and producers. Faced with severe pest problems that could not be addressed economically with pesticides, as well as

government policies that encouraged reduction of pesticide use, greenhouse plant production systems were developed based on the compatible use of natural enemies and other control tactics for the suppression of several arthropod pests (van Lenteren, 2000).

9.2.2 Trichogramma releases for suppression of stalk boring pests In Latin America and China, several million hectares of corn, sugarcane, cotton and cereals are grown under IPM systems that use augmentative releases of egg parasitoids (Trichogramma spp. and Telenomus remus) for suppression of lepidopteran pests (VanDriesche & Bellows, 1996; van Lenteren & Bueno, 2003). Because of the relatively high cost of insecticides, several governmental agencies have become actively involved in the mass production and release of natural enemies. Recently, augmentative releases of Trichogramma have targeted the European corn borer in sweet corn and pepper systems in the eastern USA. Early season releases of T. ostriniae in sweet corn fields in New York produced season long parasitism of corn borer eggs (Hoffmann et al., 2002). The efficacy of Trichogramma releases has typically been assessed by increased rates of egg parasitism and/or reduced levels of damage caused by the target pest. Smith (1996) summarized over 40 studies that examined the efficacy of releases of Trichogramma against lepidopteran pests; wide variation has been reported in the increase of egg parasitism and the percentage reduction in damage following releases.

9.2.3 Augmentative releases of insect pathogens Viruses During the 1970s Brazilian farmers began to use suspensions of diseased caterpillars to reduce densities of the velvet bean caterpillar (Anticarsia gemmatalis), a major lepidopteran pest of soybeans. These suspensions contained a nuclear polyhedrosis virus (AgNPV), applied as an inoculative microbial control, which spread to cause epizootics and suppressed populations of velvet bean caterpillar. Following a decade of research the Brazilian agricultural research organization Empresa Brasileira de Pesquisa Agropecuaria (EMBRAPA),


formulated AgNPV as a wettable powder that was used by farmer cooperatives, individual soybean farmers and private companies. In 1982, AgNPV was applied to about 2000 hectares of soybeans; by 1993 this virus was applied to over 1 million hectares of soybeans in Brazil (Moscardi, 1999). The use of this virus has resulted in significant reductions in insecticide applied to soybeans and is one of the best examples of augmentative microbial control of an insect pest. Bacteria For over four decades, several subspecies of Bacillus thuringiensis (Bt) have been widely used in inundative biological control programs for suppression of lepidopteran, dipteran and coleopteran pests (Federici, 1999). Bt is commercially available and has been applied for suppression of pests in a wide range of systems (e.g. forests, organic vegetables and aquatic environments). Bt is used within an inundative strategy for microbial control and results in rapid mortality of larval stages with no reproduction in the field. The specificity and short residual properties of Bt have been viewed as either an advantage or disadvantage of the use of Bt. Federici (1999) summarizes several factors that have contributed to the successful augmentative use of Bt: the ability to mass produce and apply on a large scale, relatively rapid mortality of target pests, activity against a number of important pest species, and fewer non-target effects and negative environmental effects compared to many insecticides. Nematodes Several species and strains (geographic isolates) of entomopathogenic nematodes in the families Steinernematidae and Heterorhabditidae are commercially produced for augmentative releases against insect pests in soil or protected habitats (Kaya & Gaugler, 1993). Nematodes used in biological control contain mutualistic bacteria (Xenorhabdus spp.), which are injected into an insect host by the infective stages of these nematodes and rapidly kill the host. These nematodes are generally considered to have broad host ranges, but recent studies have shown that several species are restricted to particular habitats and may only attack specific host taxa within that habitat. Based

on this new knowledge, previous failed attempts to use entomopathogenic nematodes in augmentative programs should be re-examined. Releases of entomopathogenic nematodes have provided biological control of numerous insect pests including the black vine weevil and citrus root weevil (Diaprepes abbreviatus) (Georgis et al., 2006).

9.2.4 Perspectives on augmentation Due to the repetitive nature of augmentative releases, comparisons with chemical control are inevitable. One of the recurring themes for augmentative releases is their relatively higher cost and lower effectiveness compared to the use of chemical insecticides (Collier & Van Steenwyk, 2004). These comparisons are based on the premise that augmentative releases can be substituted for insecticide applications, which cause immediate and close to 100% mortality. Using augmentative releases as part of an IPM program requires a change in perspective as natural enemies are not pesticides and in general they do not cause immediate mortality. The appropriate measures to determine the success of augmentative biological control should include relevant ecological and economic data from comparative field trials and an assessment of the numbers of growers using the approach, as well as the hectarage under augmentative biological control (van Lenteren, 2006). These measures will require new research foci and collaboration among producers of natural enemies, biological control specialists, extension educators and end-users. However, the case studies we have presented clearly show that such an effort will be worth it.

9.3 Conclusions Adding natural enemies to control pests, whether within a program of classical biological control or augmenting commercially available species has a long track record of success. When done properly these methods provide economic control without substantial environmental damage. Success in adding natural enemies is often restricted not by the attributes of the crop–pest system or natural enemy involved, but by self-imposed limitations based on theoretical constructs, perceptions




of what farmers want or limited experience using natural enemies. To expand the use of natural enemies in IPM will require a commitment to initiate research and extension programming when new pests are identified (as in the case of introduced pests) or as key elements of IPM programs for established or native pests. Critical research areas include increased efforts in natural enemy systematics, non-target research in the area of origin and development of protocols to maximize the effectiveness of released natural enemies. In augmentative programs, quality control and validation of efficacy are critical research areas. For both types of biological control, extension programming is needed to provide endusers information on natural enemy identification, options for use and testing compatibility of natural enemies with other control tactics. With growing global trade, shifts in market demands and increased environmental awareness, using natural enemies will become an ever-increasing component of IPM programs.

References Barbosa, P. (ed.) (1998). Conservation Biological Control. San Diego, CA: Academic Press. Bellows, T. S. Jr. & Fisher, T. W. (eds.) (1999). Handbook of Biological Control. San Diego, CA: Academic Press. Collier, T. & Van Steenwyk, R. (2004). A critical evaluation of augmentative biological control. Biological Control, 31, 245–256. Coombs, E. M., Clark, J. K., Piper, G. L. & Cofrancesco, A. F. Jr. (eds.) (2004). Biological Control of Invasive Plants in the United States. Corvallis, OR: Oregon State University Press. Debach, P. & Rosen, D. (1991). Biological Control by Natural Enemies. New York: Cambridge University Press. Delfosse, E. S. (2004). Introduction. In Biological Control of Invasive Plants in the United States, eds. E. M. Coombs, J. K. Clark, G. L. Piper & A. F. Cofrancesco, Jr., pp. 1–11. Corvallis, OR: Oregon State University Press. Federici, B. A. (1999). A perspective on pathogens as biological control agents for insect pests. In Handbook of Biological Control, eds. T. S. Bellows, Jr. & T. W. Fisher, pp. 517–548. San Diego, CA: Academic Press. Follett, P. A. & Duan, J. J. (eds.) (2000). Nontarget Effects of Biological Control. Norwell, MA: Kluwer. Georgis, R., Koppenhofer, A. M., Lacey, L. A. et al. (2006). Successes and failures in the use of parasitic nematodes for pest control. Biological Control, 38, 103–123.

Gutierrez, A. P., Caltagirone, L. E. & Meikle, W. (1999). Evaluation of results: economics of biological control. In Handbook of Biological Control, eds. T. S. Bellows, Jr. & T. W. Fisher, pp. 243–252. San Diego, CA: Academic Press. Heimpel, G. E., Ragsdale, D. W., Venette, R. et al. (2004). Prospects for importation biological control of the soybean aphid: anticipating potential costs and benefits. Annals of the Entomological Society of America, 97, 249–258. Hoffmann, M. P., Wright, M. G., Pilcher, S. A. & Gardner, J. (2002). Inoculative releases of Trichogramma ostriniae for suppression of Ostrinia nubilalis (European corn borer) in sweet corn: field biology and population dynamics. Biological Control, 25, 249–258. Hunter, C. D. (1997). Suppliers of Beneficial Organisms in North America. Sacramento, CA: California Environmental Protection Agency, Department of Pesticide Regulation, Environmental Monitoring and Pest Management Branch. Available at www. Kaya, H. K. & Gaugler, R. (1993). Entomopathogenic nematodes. Annual Review of Entomology, 38, 181–206. Moscardi, F. (1999). Assessment of the application of baculoviruses for control of Lepidoptera. Annual Review of Entomology, 44, 257–289. Neuenschwander, P. & Herren, H. R. (1988). Biological control of the cassava mealybug, Phenacoccus manihoti, by the exotic parasitoid, Epidinocaris lopezi, in Africa. Philosophical Transactions of the Royal Society of London B, 318, 319–333. Parrella, M. P., Hansen, L. S. & van Lenteren, J. C. (1999). Glasshouse environments. In Handbook of Biological Control, eds. T. S. Bellows, Jr. & T. W. Fisher, pp. 819– 839. San Diego, CA: Academic Press. Pickett, C. H. & Pitcairn, M. J. (1999). Classical biological control of ash whitefly: factors contributing to its success in California. Biological Control, 14, 143– 158. Obrycki, J. J., Lewis, L. C. & Orr, D. B. (1997). Augmentative releases of entomophagous species in annual cropping systems. Biological Control, 10, 30–36. O’Neil, R. J., Giles, K. L., Obrycki, J. J. et al. (1998). Evaluation of the quality of four commercially available natural enemies. Biological Control, 11, 1–8. Ridgway, R. L. & Vinson, S. B. (1977). Biological Control by Augmentation of Natural Enemies. New York: Plenum Press. Shelton, A. M., Zhao, J.-Z. & Roush, R. T. (2002). Economic, ecological, food safety, and social consequences of the development of Bt transgenic plants. Annual Review of Entomology, 47, 845–881.


Smith, S. M. (1996). Biological control with Trichogramma: advances, successes, and potential of their use. Annual Review of Entomology, 41, 375–406. VanDriesche, R. & Bellows, T. S. Jr. (1996). Biological Control. New York: Chapman & Hall. van Lenteren, J. C. (2000). A greenhouse without pesticides: fact or fantasy? Crop Protection, 19, 375–384. van Lenteren, J. C. (2006). How not to evaluate augmentative biological control. Biological Control, 39, 115–118.

van Lenteren, J. C. & Bueno, V. H. P. (2003). Augmentative biological control of arthropods in Latin America. BioControl, 48, 121–139. van Lenteren, J. C., Hale, A., Klapwijk, J. N., van Schelt, J. & Stenberg, S. (2003). Guidelines for quality control of commercially produced natural enemies. In Quality Control and Production of Biological Control Agents: Theory and Testing, ed. J. C. van Lenteren, pp. 265–304. Cambridge, MA: CABI Publishing.


Chapter 10

Crop diversification strategies for pest regulation in IPM systems Miguel A. Altieri, Clara I. Nicholls and Luigi Ponti Ninety-one percent of the 1500 million hectares of the worldwide cropland are mostly under annual crop monocultures of wheat, rice, maize, cotton and soybeans (Smith & McSorley, 2000). These systems represent an extreme form of simplification of nature’s biodiversity, since monocultures, in addition to being genetically uniform and speciespoor systems, advance at the expense of natural vegetation, a key landscape component that provides important ecological services to agriculture such as natural mechanisms of crop protection (Altieri, 1999). Since the onset of agricultural modernization, farmers and researchers have been faced with a main ecological dilemma arising from the homogenization of agricultural systems: an increased vulnerability of crops to insect pests and diseases, which can be devastating when infesting uniform crop, large-scale monocultures (Adams et al., 1971; Altieri & Letourneau, 1982, 1984). The expansion of monocultures has decreased abundance and activity of natural enemies due to the removal of critical food resources and overwintering sites (Corbett & Rosenheim, 1996). With accelerating rates of habitat removal, the contribution to pest suppression by biocontrol agents using these habitats is declining and consequently agroecosystems are becoming increasingly vulnerable to pest outbreaks (e.g. Gurr et al., 2004).

A key task for agroecologists is to understand the link between biodiversity reduction and pest incidence in modern agroecosystems in order to reverse such vulnerability by increasing functional diversity in agricultural landscapes. One of the most obvious advantages of diversification is a reduced risk of total crop failure due to pest infestations (Nicholls & Altieri, 2004). Research results from recent vineyard studies conducted in California illustrate ways in which biodiversity can contribute to the design of pest-stable agroecosystems by creating an appropriate ecological infrastructure within and around cropping systems.

10.1 Understanding pest vulnerability in monocultures The spread of modern agriculture has resulted in tremendous changes in landscape diversity. There has been a consistent trend toward simplification that entails: (1) enlargement of fields, (2) aggregation of fields, (3) increase in the density of crop plants, (4) increase in the uniformity of crop population age structure and physical quality and (5) decrease in inter- and intraspecific diversity within the planted field. One of the main characteristics of the modern agricultural landscape is

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


the large size and homogeneity of crop monocultures, which fragment the natural landscape. This can directly affect the abundance and diversity of natural enemies, as the larger the area under monoculture the lower the viability of a given population of beneficial fauna. Moreover, monocultures do not constitute good environments for natural enemies (Andow, 1991). Such simple crop systems lack many of the resources such as refuge sites, pollen, nectar and alternative prey and hosts, that natural enemies need to feed, reproduce and thrive. To the pests, the monocrop is a dense and pure concentration of its basic food resource, so, of course, many insect herbivores boom in such fertilized, weeded and watered fields. For the natural enemies, such overly simplified cropping systems are less hospitable because natural enemies require more than prey/hosts to complete their life cycles. Many parasitoid adults, for instance, require pollen and nectar to sustain themselves while they search for hosts (Gurr et al., 2004). Normal cultural activities such as tillage, weeding, spraying and fertilization can have serious effects on farm insects. Insect pest outbreaks and/or resurgences following insecticide applications are common phenomena (Pimentel & Perkins, 1980). Pesticides either fail to control the target pests or create new pest problems. Development of resistance in insect pest populations is the main way in which pesticide use can lead to pest control failure. Pesticides can also foster pest outbreaks through the elimination of the target pest’s natural enemies (Pimentel & Lehman, 1993). Research also suggests that the susceptibility of crops to insects may be affected by the application of chemical fertilizers. Studies show that increases in nitrogen rates dramatically increased aphid and mite numbers. In reviewing 50 years of research relating to crop nutrition and insect attack, Scriber (1984) found 135 studies showing increased damage and/or growth of leaf-chewing insects or mites in nitrogen-fertilized crops, versus fewer than 50 studies in which herbivore damage was reduced by normal fertilization regimes. Pest levels were highly correlated to increased levels of soluble nitrogen in leaf tissue, suggesting that high nitrogen inputs can precipitate high levels of herbivore damage in crops (Altieri & Nicholls, 2003).

Lately, research suggests that transgenic crops, now ubiquitous components of agricultural landscapes, may affect natural enemy species directly through inter-trophic-level effects of the toxin. The potential for Bt toxins moving through arthropod food chains poses serious implications for natural biocontrol in agricultural fields. Studies show that the Bt toxin can affect beneficial insect predators that feed on insect pests present in Bt crops (Hilbeck et al., 1998). Inter-trophic-level effects of the Bt toxin raise serious concerns about the potential of the disruption of natural pest control (Altieri, 2000).

10.2 Biodiversity in agroecosystems: types and roles Biodiversity refers to all species of plants, animals and microorganisms existing and interacting within an ecosystem, and which play important ecological functions such as pollination, organic matter decomposition, predation or parasitism of undesirable organisms and detoxification of noxious chemicals (Gliessman, 1998). These renewal processes and ecosystem services are largely biological; therefore their persistence depends upon maintenance of ecological diversity and integrity. When these natural services are lost due to biological simplification, the economic and environmental costs can be quite significant. Economically, in agriculture the burdens include the need to supply crops with costly external inputs, since agroecosystems deprived of basic regulating functional components lack the capacity to sponsor their own soil fertility and pest regulation (Conway & Pretty, 1991). The biodiversity components of agroecosystems can be classified in relation to the role they play in the functioning of cropping systems. According to this, agricultural biodiversity can be grouped as follows (Altieri, 1994; Gliessman, 1998).

r Productive biota: crops, trees and animals chosen by farmers that play a determining role in the diversity and complexity of the agroecosystem.




Fig. 10.1 The relationship between planned biodiversity (which the farmer determines, based on management of the agroecosystem) and associated biodiversity and how the two promote ecosystem function (modified from Vandermeer & Perefecto, 1995).

r Resource biota: organisms that contribute to productivity through pollination, biological control, decomposition, etc. r Destructive biota: weeds, insects pests, microbial pathogens, etc., which farmers aim at reducing through cultural management. The above categories of biodiversity can further be recognized as two distinct components (Vandermeer & Perefecto, 1995). The first component, planned biodiversity, includes the crops and livestock purposely included in the agroecosystem by the farmer, which will vary depending on the management inputs and crop spatial/temporal arrangements. The second component, associated biodiversity, includes all soil flora and fauna, herbivores, carnivores, decomposers, etc., that colonize the agroecosystem from surrounding environments and that will thrive in the agroecosystem depending on its management and structure. The relationship of both types of biodiversity components is illustrated in Fig. 10.1. Planned biodiversity has a direct function, as illustrated by the bold arrow connecting the planned biodiversity box with the ecosystem function box. Associated biodiversity also has a function, but it is mediated through planned biodiversity. Thus, planned biodiversity also has an indirect function, illustrated by the dotted arrow in the figure, which is realized through its influence on the associated biodi-

versity. For example, the trees in an agroforestry system create shade, which makes it possible to grow only sun-intolerant crops. So, the direct function of this second species (the trees) is to create shade. Yet along with the trees might come wasps that seek out the nectar in the tree’s flowers. These wasps may in turn be the natural parasitoids of pests that normally attack understory crops. The wasps are part of the associated biodiversity. The trees, then create shade (direct function) and attract wasps (indirect function) (Vandermeer & Perefecto, 1995). The optimal behavior of agroecosystems depends on the level of interactions between the various biotic and abiotic components. By assembling a functional biodiversity it is possible to initiate synergisms which subsidize agroecosystem processes by providing ecological services such as the activation of soil biology, the recycling of nutrients, the enhancement of beneficial arthropods and antagonists, and so on, all important components that determine the sustainability of agroecosystems (Nicholls et al., 2000). The key is to identify the type of biodiversity that it is desirable to maintain and/or enhance in order to carry out ecological services, and then to determine the best practices that will encourage the desired biodiversity components. There are many agricultural practices and designs that have the potential to enhance functional


Fig. 10.2 Effects of agroecosystem management and associated cultural practices on biodiversity of natural enemies and abundance of insect pests.


hedgerows shelterbelts windbreaks



Habitat diversification

cover crops

Organic soil management

Low soil disturbance tillage practices


Cultural practices

Conventional tillage

Total weed removal



Chemical fertilization

Decrease in Natural Enemies Species Diversity Population Increases of Pestiferous Species

biodiversity, and others that negatively affect it (Fig. 10.2). The idea is to apply the best management practices in order to enhance or regenerate the kind of biodiversity that can best subsidize the sustainability of agroecosystems by providing ecological services such as biological pest control, nutrient cycling, water and soil conservation, etc.

10.3 Diversified agroecosystems and pest management Across the world, agroecosystems differ in age, diversity, structure and management. In fact, there is great variability in basic ecological and agronomic patterns among the various dominant agroecosystems. In general, agroecosystems that

are more diverse, more permanent, isolated, and managed with low input technology (e.g. agroforestry systems, traditional polycultures) take fuller advantage of work usually done by ecological processes associated with higher biodiversity than highly simplified, input-driven and disturbed systems (e.g. modern vegetable monocultures and orchards). All agroecosystems are dynamic and subjected to different levels of management so that the crop arrangements in time and space are continually changing in the face of biological, cultural, socioeconomic and environmental factors. Such landscape variations determine the degree of spatial and temporal heterogeneity characteristic of agricultural regions, which may or may not benefit the pest protection of particular agroecosystems. Thus, one of the main challenges facing agroecologists today is identifying the types of




Annual crop based

Perennial crop based

Tropical homegardens




Cover crops

Mixed cropping


Cover crops


Legume based



Grains Row crops

Relay cropping Monolayered

Monocultures Monocultures

Strip cropping

Non-legume based


Decreasing level of biodiversity Increasing possibility for pest buildup Fig. 10.3 A classification of dominant agricultural agroecosystems on a gradient of diversity and vulnerability to pest outbreak.

heterogeneity (either at the field or regional level) that will yield desirable agricultural results (i.e. pest regulation), given the unique environment and entomofauna of each area. This challenge can only be met by further analyzing the relationship between vegetation diversification and the population dynamics of herbivore species, in light of the diversity and complexity of site-specific agricultural systems. A hypothetical pattern in pest regulation according to agroecosystem temporal and spatial diversity is depicted in Fig. 10.3. According to this “increasing probability for pest buildup” gradient, agroecosystems on the left side of the gradient are more biodiverse, and tend to be more amenable to manipulation since polycultures already contain many of the key factors required by natural enemies. There are, however, habitat manipulations that can introduce appropriate diversity into the important (but biodiversity impoverished) grain, vegetable and row crop systems lying in the right half of Fig. 10.3. Although herbivores vary widely in their response to crop distribution, abundance and dispersion, the majority of agroecological studies show that structural (i.e. spatial and temporal crop arrangement) and management (i.e. crop diver-

sity, input levels, etc.) attributes of agroecosystems influence herbivore dynamics. Several of these attributes are related to biodiversity and most are amenable to management (i.e. crop sequences and associations, weed diversity, genetic diversity, etc.). Diversified cropping systems, such as those based on intercropping and agroforestry or cover cropping of orchards, have been the target of much research recently. This interest is largely based on the emerging evidence that these systems are more stable and more resource conserving (Vandermeer, 1995). Many of these attributes are connected to the higher levels of functional biodiversity associated with complex farming systems. As diversity increases, so do opportunities for coexistence and beneficial interference between species that can enhance agroecosystem sustainability (Vandermeer, 1995). Diverse systems encourage complex food webs which entail more potential connections and interactions among members, and many alternative paths of energy and material flow through it. For this and other reasons a more complex community exhibits more stable production and less fluctuations in the numbers of undesirable organisms. Studies further suggest that the more diverse the agroecosystems and the longer this diversity remains undisturbed, the more internal links develop to promote greater insect stability. It is clear, however, that the stability of the insect


community depends not only on its trophic diversity, but also on the actual density-dependence nature of the trophic levels (Southwood & Way, 1970). In other words, stability will depend on the precision of the response of any particular trophic link to an increase in the population at a lower level. What is apparent is that functional characteristics of component species are as important as the total number of species in determining processes and services in ecosystems (Tilman, 1996). From an agroecosystem management point of view, the focus should be placed on enhancing a specific biodiversity component that plays a specific role, such as a plant that fixes nitrogen, provides cover for soil protection or harbors resources for natural enemies. In the case of farmers without major economic and resource limits and who can withstand a certain risk of crop failure, a crop rotation or a simple polyculture may be all it takes to achieve a desired level of stability. But in the case of resource-poor farmers, who can not tolerate crop failure, highly diverse cropping systems would probably be the best choice. The obvious reason is that the benefit of complex agroecosystems is low risk; if a species falls to disease, pest attack or weather, another species is available to fill the void and maintain full use of resources. Thus there are potential ecological benefits to having several species in an agroecosystem: compensatory growth, full use of resources and nutrients and pest protection (Ewel, 1999).

10.4 Diversity in traditional farming systems The persistence of millions of hectares under traditional agriculture in the form of polycultures, agroforestry systems, etc. documents a successful indigenous agricultural biodiversification strategy for adapting to difficult environments (Altieri, 1999). These microcosms of traditional agriculture offer promising models for other areas as they promote biodiversity, thrive without agrochemicals and sustain year-round yields (Denevan, 1995). Traditional crop management practices used by many

resource-poor farmers represent a rich resource for agroecologists interested in understanding the mechanisms at work in complex agroecosystems, especially the interactions between biodiversity and ecosystem function. Some agroecologists have recognized the virtues of diversified traditional agroecosystems whose sustainability lies in the complex ecological models they follow, quickly realizing that the prevalence of these systems is of key importance to peasants, as interactions between crops, animals and trees result in beneficial synergisms allowing agroecosystems to sponsor their own soil fertility, pest control and productivity (Altieri et al., 1985; Reinjtes et al., 1992). Considerable work was conducted on the biological mechanisms at play within intercropping systems which minimize crop losses due to insect pests, diseases and weeds (Altieri, 1994). The literature is full of examples of experiments documenting that diversified traditional cropping systems such as the prevalent maize–bean polyculture of the Latin American tropics exhibit reduced pest populations (Andow, 1991; Landis et al., 2000; Altieri & Nicholls, 2004). Researchers have shown that it is only when traditional systems are modernized, reducing their plant diversity, that herbivore abundance increases to pest levels, compounded by changes brought about by modern plant breeding and agronomy. In fact, although traditional farmers may be aware that insects can exert crop damage, they rarely consider them pests, as experienced by Morales et al. (2001) when studying traditional methods of pest control among the highland Maya of Guatemala. Influenced by Mayan attitudes, these Western scientists rapidly reformulated their research questions and rather than study how Mayan farmers control pest problems, they focused on why farmers do not have pest problems. This line of inquiry proved more productive as it allowed researchers to understand how farmers designed and managed pest-resilient cropping systems and to explore the mechanisms underlying agroecosystem health. Many studies have transcended the research phase and have found applicability to control specific pests such as lepidopteran stem borers in Africa. Scientists at the International Center of Insect Physiology and Ecology (ICIPE)




developed a habitat management system which uses plants in the borders of maize fields which act as trap crops (Napier grass and Sudan grass) attracting stem borer colonization away from maize (the push) and two plants intercropped with maize (molasses grass and silverleaf) that repel the stem borers (the pull) (Khan et al., 2000). Border grasses also enhance the parasitization of stem borers by the wasp Cotesia semamiae and are important fodder plants. The leguminous silverleaf (Desmodium uncinatum) suppresses the parasitic weed Striga spp. by a factor of 40 when compared with maize monocrop. Desmodium’s nitrogen-fixing ability increases soil fertility; and it is an excellent forage. As an added bonus, sale of Desmodium seed is proving to be a new income-generating opportunity for women in the project areas. The push–pull system has been tested on over 450 farms in two districts of Kenya and has now been released for uptake by the national extension systems in East Africa. Participating farmers in the breadbasket of Trans Nzoia are reporting a 15–20% increase in maize yield. In the semi-arid Suba district – plagued by both stem borers and Striga – a substantial increase in milk yield has occurred in the last four years, with farmers now being able to support increased numbers of dairy cows on the fodder produced. When farmers plant maize together with the push–pull plants, a return of $US 2.30 for every dollar invested is made, as compared to only $US 1.40 obtained by planting maize as a monocrop (Khan et al., 2000).

10.5 Plant diversity and insect pest incidence An increasing body of literature documents that increased plant diversity in agroecosystems leads to pest population regulation by favoring the abundance and efficacy of associated natural enemies (Landis et al., 2000). Research has shown that mixing certain plant species usually leads to density reductions of specialized herbivores. In a review of 150 published investigations Risch et al. (1983) found evidence to support the notion that specialized insect herbivores were less numerous

in diverse systems (53% of 198 cases). In another comprehensive review of 209 published studies that deal with the effects of vegetation diversity in agroecosystems on herbivores arthropod species, Andow (1991) found that 52% of the 287 total herbivore species examined in these studies were less abundant in polycultures than in monocultures, while only 15.3% (44 species) exhibited higher densities in polycultures. In a more recent review of 287 cases, Helenius (1998) found that the reduction of monophagous pests was greater in perennial systems, and that the reduction of polyphagous pest numbers was less in perennial than in annual systems. Helenius (1998) concluded that monophagous (specialist) insects are more susceptible to crop diversity than polyphagous insects. He cautioned about the increased risk of pest attack if the dominant herbivore fauna in a given agroecosystem is polyphagous. In his analysis of available studies on crop–weed systems, Baliddawa (1985) found that 56% of pest reductions in weed diversified cropping systems were caused by natural enemies. In examining numerous studies testing the responses of pest and beneficial arthropods to plant diversification in cruciferous crops, Hooks & Johnson (2003) concluded that biological parameters of herbivores impacted by crop diversification were mainly related to the behavior of the insect studied. Mechanisms accounting for herbivore responses to plant mixtures include reduced colonization, reduced adult tenure time in the crop and oviposition interference. The ecological theory relating to the benefits of mixed versus simple cropping systems revolves around two possible explanations of how insect pest populations attain higher levels in monoculture systems compared with diverse ones. The two hypotheses proposed by Root (1973) are: The enemies hypothesis which argues that pest numbers are reduced in more diverse systems because the activity of natural enemies is enhanced by environmental opportunities prevalent in complex systems; The resource concentration hypothesis argues that the presence of a more diverse flora has direct negative effects on the ability of the insect pest to find and utilize its host plant and also to remain in the crop habitat.


The resource concentration hypothesis predicts lower pest abundance in diverse communities because a specialist feeder is less likely to find its host plant due to the presence of confusing masking chemical stimuli, physical barriers to movement or other environmental effects such as shading. It will tend to remain in the intercrop for a shorter period of time simply because the probability of landing on a non-host plant is increased; it may have a lower survivorship and/or fecundity (Bach, 1980). The extent to which these factors operate will depend on the number of host plant species present and the relative preference of the pest for each, the absolute density and spatial arrangement of each host species and the interference effects from more host plants. The enemies hypothesis attributes lower pest abundance in intercropped or more diverse systems to a higher density of predators and parasitoids (Bach, 1980). The greater density of natural enemies is caused by an improvement in conditions for their survival and reproduction, such as a greater temporal and spatial distribution of nectar and pollen sources, which can increase parasitoid reproductive potential and abundance of alternative host/prey when the pest species are scarce or at inappropriate stages (Risch, 1981). In theory, these factors can combine to provide more favorable conditions for natural enemies and thereby enhance their numbers and effectiveness as control agents.

10.6 Designing biodiverse pest-suppressive agroecosystems In real situations, exploiting the complementarity and synergy that result from the various spatial and temporal polycultural combinations involves agroecosystem design and management and requires an understanding of the numerous relationships among soils, microorganisms, plants, insect herbivores and natural enemies. Different options to diversify cropping systems are available depending on whether the current monoculture systems to be modified are based on annual or perennial crops. Diversification can

also take place outside the farm, for example, in crop-field boundaries with windbreaks, shelterbelts and living fences, which can improve habitat for beneficial insects (Altieri & Letourneau, 1982). When done correctly, plant diversification creates a suitable ecological infrastructure within the agricultural landscape providing key resources and habitat for natural enemies. These resources must be integrated into the landscape in a way that is spatially and temporally favorable to natural enemies and practical for farmers to implement. During the last decade we have applied biodiversification strategies to the design and management of pest suppressive organic vineyards in northern California. Results from some of our previously published studies (Nicholls et al., 2000, 2001, 2005), are presented here in an effort to systematize the emerging lessons from our experience on biodiversity enhancement for ecologically based pest management in agroecosystems.

10.6.1 Vineyard studies Our studies took advantage of an organic vineyard located in Mendocino County, California in which a 600-m corridor composed by 65 flowering species connected to a riparian forest cutting across the monoculture organic vineyard. This setting allowed for testing the idea whether such a corridor served as a biological highway for the movement and dispersal of natural enemies into the center of the vineyard (Nicholls et al., 2001). We evaluated whether the corridor acted as a consistent, abundant and well-dispersed source of alternative food and habitat for a diverse community of generalist predators and parasitoids, allowing predator and parasitoid populations to develop in the area of influence of the corridor well in advance of vineyard pest populations. The corridor would serve as a conduit for the dispersion of predators and parasitoids within the vineyard, thus providing protection against insect pests within the area of influence of the corridor, by allowing distribution of natural enemies throughout the vineyard. As the vineyard also contained summer cover crops, we hypothesized that neutral insects (non-pestiferous herbivores) and pollen and nectar in the summer cover crops provide a constant and abundant supply of food




Fig. 10.4 Densities of adult Erythroneura elegantula leafhoppers in cover cropped and monoculture vineyards in Mendocino County, California, during the 1996 and 1997 growing seasons.

sources for natural enemies. This in turn decouples predators and parasitoids from a strict dependence on grape herbivores, allowing natural enemies to build up in the system, thereby keeping pest populations at acceptable levels. We also conducted research at a 17-ha biodynamic vineyard located in Sonoma County, California. As part of a whole-farm biodiversity management strategy, a 0.5-ha island of flowering shrubs and herbs (insectory) was created at the center of the vineyard. This insectory was planted to provide flower resources from early April to late September to beneficial organisms, including natural enemies of grape insect pests (Nicholls et al., 2005).

10.6.2 Enhancing within vineyard biodiversity with cover crops Because most farmers either mow or plow under cover crops in the late spring, organic vineyards become virtual monocultures without floral diversity in early summer. Maintaining a green cover during the entire growing season is crucial to provide habitat and alternate food for natural enemies. An approach to achieve this is to sow summer cover crops that bloom early and throughout the season, thus providing a highly consistent, abundant and well-dispersed alternative food source, as well as microhabitats, for a diverse community of natural enemies (Nicholls et al., 2000). Maintaining floral diversity throughout the growing season in the Mendocino vineyard in the

form of summer cover crops of buckwheat and sunflower, substantially reduced the abundance of western grape leafhopper (Erythroneura elegantula) and western flower thrips (Frankliniella occidentalis) by allowing an early buildup of natural enemies. In two consecutive years (1996–1997), vineyard systems with flowering cover crops were characterized by lower densities of leafhopper nymphs and adults (Fig. 10.4). Thrips also exhibited reduced densities in vineyards with cover crops in both seasons (Nicholls et al., 2000). During both years, general predator populations on the vines were higher in the covercropped sections than in the monocultures. Generally, the populations were low early in the season and increased as prey became more numerous as the season progressed. Dominant predators included spiders, Nabis sp., Orius sp., Geocoris sp., Coccinellidae and Chrysoperla spp. Although Anagrus epos, the most important leafhopper parasitoid wasp, achieved high numbers and inflicted noticeable mortality of grape leafhopper eggs, this impact was not substantial enough. Apparently the wasps encountered sufficient food resources in the cover crops, and few moved to the vines to search for leafhopper eggs. For this reason, cover crops were mowed every other row to force movement of Anagrus wasps and predators into the vines. Before mowing, leafhopper nymphal densities on vines were similar in the selected cover-cropped rows. One week after mowing, numbers of nymphs declined on vines





25 Mean no. nymphs/leaf

Fig. 10.5 (A) Effect of cover-crop mowing in vineyards on densities of leafhopper nymphs during the 1997 growing season in Mendocino County, California. (B) Effect of cover-crop mowing in vineyards on densities of Anagrus epos during the 1997 growing season in Mendocino County, California.


20 No mow Mow 15



0 Before mow

1 week after

2 weeks after B


Mean no. Anagrus/yellow sticky trap

7000 6000 5000 no mow mow

4000 3000 2000 1000 0 Before mow

where the cover crop was mowed, coinciding with an increase in Anagrus densities in mowed covercrop rows. During the second week, such nymphal decline was even more pronounced, coinciding with an increase in numbers of Anagrus wasps in the foliage (Fig. 10.5). The mowing experiment suggests a direct ecological linkage, as the cutting of the cover crop vegetation forced the movement of Anagrus wasps and other predators harbored by the flowers, resulting in both years in a decline of leafhopper num-

1 week after

2 weeks after

bers on the vines adjacent to the mowed cover crops. Obviously, the timing of mowing must take place when eggs are present on the vine leaves in order to optimize the efficiency of arriving Anagrus wasps.

10.6.3 Corridor influences on population gradients of leafhoppers, thrips and associated natural enemies Studies assessing the influence of adjacent vegetation or natural enemy refuges on pest dynamics



Fig. 10.6 Seasonal patterns of adult leafhoppers in vineyard near and far from the corridor (Mendocino County, California, 1996).

within vineyards show that, in the case of prune refuges, the effect is limited to only a few vine rows downwind, as the abundance of A. epos exhibited a gradual decline in vineyards with increasing distance from the refuge (Corbett & Rosenheim, 1996). This finding poses an important limitation to the use of prune trees, as the colonization of grapes by A. epos is limited to field borders, leaving the central rows of the vineyard void of biological control protection. The corridor connected to a riparian forest and cutting across the vineyard was established to overcome this limitation (Nicholls et al., 2001). The flowering sequence of the various plant species provided a continual source of pollen and nectar, as well as a rich and abundant supply of neutral insects for the various predator species, thus allowing the permanence and circulation of viable populations of key species within the corridor. In both years, adult leafhoppers exhibited a clear density gradient, with lowest numbers in vine rows near the corridor and increasing numbers towards the center of the field. The highest concentration of adult and nymph leafhoppers occurred after the first 20–25 rows (30–40 m) downwind from the corridor (Fig. 10.6). A similar population and distribution gradient was apparent for thrips. In both years, catches of leafhoppers and thrips were substantially higher in the central rows than in rows adjacent to the corridor.

The abundance and spatial distribution of generalist predators in the families Coccinellidae, Chrysopidae, Nabidae and Syrphidae were influenced by the presence of the corridor which channeled dispersal of the insects into adjacent vines (Fig. 10.7). Predator numbers were higher in the first 25 m adjacent to the corridor, which probably explains the reduction of leafhoppers and thrips observed in the first 25 m of vine rows near the corridor. The presence of the corridor was associated with the early vineyard colonization by Anagrus wasps, but this did not result in a net seasonlong prevalence in leafhopper egg parasitism rates in rows adjacent to the corridor. The proportion of eggs parasitized tended to be uniformly distributed across all rows in both blocks. Eggs in the center rows had slightly higher mean parasitization rates than eggs located in rows near the corridor, although differences were not statistically significant.

10.6.4 Creating flowering islands as a push–pull system for natural enemies in a Sonoma vineyard Cover crops and corridors are all important practices to enhance insect biodiversity, but at times creating habitat on less productive parts of the farm to concentrate natural enemies may be a key strategy. This is the approach used at Benziger farm in Sonoma County, where a 0.25-ha island of flowering shrubs and herbs was created at the


Fig. 10.7 Seasonal patterns of predator catches in vineyard as influenced by the presence or absence of forest edge and the corridor (P < 0.05; Mann–Whitney U-test) (Mendocino County, California, 1996).

center of the vineyard to act as a push–pull system for natural enemy species (Nicholls et al., 2005). The island and its mix of shrubs and herbs provides flower resources from early April to late September to a number of herbivore insects (pests, neutral non-pestiferous insects and pollinators) and associated natural enemies which build up in the habitat, later dispersing into the vineyard. The island acts as a source of pollen, nectar and neutral insects which serve as alternate food to a variety of predators and parasites including Anagrus wasps. The island is dominated by neutral insects that forage on the various plants but also serve as food to natural enemies of thrips which slowly build up in numbers in the adjacent vineyard as the season progresses. Many natural enemies moved from this island insectory into the vineyard (up to 60 m). Responding to the abundance of habitat resources in the insectory, predators tended to decrease in abundance in vines 30 and 60 m away. Orius spp. reached significantly lower abundances in vines away from the insectory a trend that correlated with the densities of thrips which increased in vines far from the island. While the proportion of natural enemies in relation to the total number of insects caught in the traps remained relatively constant within the insectarium, their proportion increased from 1% to 10% and 13% in vines located 30 m and 60 m from the insectory respectively. Orius spp. and coccinellids are prevalent colonizers at the

Table 10.1 Levels of leafhopper eggs parasitization by Anagrus wasps in the second and tenth vine rows from the insectory during the peak summer months (Sonoma County, California, 2004) Parasitization (%) Date 12 July 26 July 23 August

Second row

Tenth row

76.8 52.4 43.6

54.1 52.3 32.1

beginning of the season, but later syrphid flies and Anagrus wasps started dispersing from the insectarium into the vineyard (Fig. 10.8). Parasitization of leafhopper eggs by Anagrus wasps was particularly high on the vines near the island (10 m from the island), with parasitization levels decreasing progressively towards the center of the vineyard away from the island (Table 10.1). It is possible that the presence of pollen and nectar in the island’s flowers build up the populations of Anagrus, which moved from the island but confined their activity to nearby rows.

10.7 Conclusions The instability of agroecosystems, manifested in the form of pest outbreaks, can be




Fig. 10.8 Dispersal of Anagrus wasps and generalist predators from the island into the vineyard (Sonoma County, California, 2004).

increasingly linked to the expansion of crop monocultures (Altieri, 1994). Plant communities that are intensely modified to meet the special needs of humans become subject to heavy pest damage. The inherent self-regulation characteristics of natural communities are lost when humans modify such communities through the shattering of the fragile thread of community interactions. Agroecologists maintain that restoring the shattered elements of the community homeostasis through the addition or enhancement of biodiversity can repair this breakdown (Gliessman, 1998; Altieri, 1999). A key strategy in sustainable agriculture is to reincorporate diversity into the agricultural landscape and manage it more effectively. Emergent ecological properties develop in diversified agroecosystems that allow the system to function in ways that maintain soil fertility, crop production and pest regulation. The role of agroecolo-

gists should be to encourage agricultural practices that increase the abundance and diversity of above- and below-ground organisms, which in turn provide key ecological services to agroecosystems. The main approach in ecologically based pest management is to use management methods that increase agroecosystem diversity and complexity as a foundation for establishing beneficial interactions that keep pest populations in check (Altieri & Nicholls, 2004). This is particularly important in underdeveloped countries where resource-poor farmers have no access to sophisticated inputs and rely instead on the rich complex of predators and parasites associated with their mixed cropping systems for insect pest control. Reduction of plant diversity in such systems brought by modernization has the potential to disrupt natural pest control mechanisms, making farmers more dependent on pesticides. When properly implemented, habitat management leads to establishment of the desired type of plant biodiversity and the ecological infrastructure necessary for attaining optimal natural enemy diversity and abundance. Habitat


management may not always demand a radical change in farming as illustrated by the relative ease with which cover crops or corridors can be introduced into vineyard systems, providing a highly consistent, abundant and well-dispersed alternative food source, as well as microhabitats, for a diverse community of natural enemies, thus bringing biological control benefits to farmers (Landis et al., 2000). Long-term maintenance of diversity requires a management strategy that considers the design of environmentally sound agroecosystems above purely economic concerns. This is why several authors have repeatedly questioned whether the pest problems of modern agriculture can be ecologically alleviated within the context of the present capital-intensive structure of agriculture (Altieri & Nicholls, 2004; Gurr et al., 2004). Many problems of modern agriculture are rooted within that structure and thus require consideration of major social change, land reform, redesign of machinery and ecological reorientation of research and extension to increase the possibilities of improved pest control through vegetation management. Whether the potential and spread of ecologically based pest management is realized will depend on policies, attitude changes on the part of researchers and policy makers, existence of markets for organic produce, and also the organization of farmer and consumer movements that demand a more healthy and viable agriculture.

References Adams, M. W., Ellingbae, A. H. & Rossineau, E. C. (1971). Biological uniformity and disease epidemics. BioScience, 21, 1067–1070. Altieri, M. A. (1994). Biodiversity and Pest Management in Agroecosystems. New York: Haworth Press. Altieri, M. A. (1999). The ecological role of biodiversity in agroecosystems. Agriculture Ecosystems & Environment, 74, 19–31. Altieri, M. A. (2000). The ecological impacts of transgenic crops on agroecosystem health. Ecosystem Health, 6, 13– 23. Altieri, M. A. & Letourneau, D. K. (1982). Vegetation management and biological control in agroecosystems. Crop Protection, 1, 405–430.

Altieri, M. A. & Letourneau, D. K. (1984). Vegetation diversity and outbreaks of insect pests. CRC Critical Reviews in Plant Sciences, 2, 131–169. Altieri, M. A. & Nicholls, C. I. (2003). Soil fertility management and insect pests: harmonizing soil and plant health in agroecosystems. Soil and Tillage Research, 72, 203. Altieri, M. A. & Nicholls, C. I. (2004). Biodiversity and Pest Management in Agroecosystems, 2nd edn. New York: Haworth Press. Altieri, M. A., Wilson, R. C. & Schmidt, L. L. (1985). The effects of living mulches and weed cover on the dynamics of foliage–arthropod and soil–arthropod communities in 3 crop systems. Crop Protection, 4, 201– 213. Andow, D. A. (1991). Vegetational diversity and arthropod population response. Annual Review of Entomology, 36, 561–586. Bach, C. E. (1980). Effects of plant diversity and time of colonization on an herbivore–plant interaction. Oecologia, 44, 319–326. Baliddawa, C. W. (1985). Plant species diversity and crop pest control: an analytical review. Insect Science and Its Applications, 6, 479–487. Conway, G. R. & Pretty, J. (1991). Unwelcome Harvest: Agriculture and Pollution. London: Earthscan. Corbett, A. & Rosenheim, J. A. (1996). Impact of a natural enemy overwintering refuge and its interaction with the surrounding landscape. Ecological Entomology, 21, 155–164. Denevan, W. M. (1995). Prehistoric agricultural methods as models for sustainability. Advances in Plant Pathology, 11, 21–43. Ewel, J. J. (1999). Natural systems as models for the design of sustainable systems of land use. Agroforestry Systems, 45, 1–21. Gliessman, S. R. (1998). Agroecology: Ecological Processes in Sustainable Agriculture. Chelsea, MI: Ann Arbor Press. Gurr, G. M., Wratten, S. D. & Altieri, M. A. (2004). Ecological Engineering for Pest Management: Advances in Habitat Manipulation for Arthropods. Wallingford, UK: CABI Publishing. Helenius, J. (1998). Enhancement of predation through within-field diversification. In Enhancing Biological Control, eds. E. Pickett & R. L. Bugg, pp. 121–160. Berkeley, CA: University of California Press. Hilbeck, A., Baumgartner, M., Fried, P. M. & Bigler, F. (1998). Effects of transgenic Bacillus thuringiensis cornfed prey on mortality and development time of immature Chrysoperla carnea (Neuroptera: Chrysopidae). Environmental Entomology, 27, 480–487.




Hooks, C. R. R. & Johnson, M. W. (2003). Impact of agricultural diversification on the insect community of cruciferous crops. Crop Protection, 22, 223–238. Khan, Z. R., Pickett, J. A., van der Berg, J. & Woodcock, C. M. (2000). Exploiting chemical ecology and species diversity: stemborer and Striga control for maize in Africa. Pest Management Science, 56, 1–6. Landis, D. A., Wratten, S. D. & Gurr, G. M. (2000). Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175–201. Morales, H., Perfecto, I. & Ferguson, B. (2001). Traditional soil fertilization and its impacts on insect pest populations in corn. Agriculture, Ecosystems and Environment, 84, 145–155. Nicholls, C. I. & Altieri, M. A. (2004). Designing speciesrich, pest-suppressive agroecosystems through habitat management. In Agroecosystems Analysis, eds. D. Rickerl & C. Francis, pp. 49–61. Madison, WI: American Society of Agronomy. Nicholls, C. I., Parrella, M. P. & Altieri, M. A. (2000). Reducing the abundance of leafhoppers and thrips in a northern California organic vineyard through maintenance of full season floral diversity with summer cover crops. Agricultural and Forest Entomology, 2, 107–113. Nicholls, C. I., Parrella, M. & Altieri, M. A. (2001). The effects of a vegetational corridor on the abundance and dispersal of insect biodiversity within a northern California organic vineyard. Landscape Ecology, 16, 133–146. Nicholls, C. I., Ponti, L. & Altieri, M. A. (2005). Manipulating vineyard biodiversity for improved insect pest management: case studies from Northern California. International Journal of Biodiversity Science and Management, 1, 191–203.

Pimentel, D. & Lehman, H. (1993). The Pesticide Question. New York: Chapman & Hall. Pimentel, D. & Perkins, J. H. (1980). Pest Control: Cultural and Environmental Aspects, AAAS Selected Symposium No. 43. Boulder, CO: Westview Press. Reinjtes, C., Haverkort, B. & Waters-Bayer, A. (1992). Farming for the Future. London: Macmillan. Risch, S. J. (1981). Insect herbivore abundance in tropical monocultures and polycultures: an experimental test of two hypotheses. Ecology, 62, 1325–1340. Risch, S. J., Andow, D. & Altieri, M. A. (1983). Agroecosystem diversity and pest control: data, tentative conclusions, and new research directions. Environmental Entomology, 12, 625–629. Root, R. (1973). Organization of a plant–arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecological Monographs, 43, 95–124. Scriber, J. M. (1984). Nitrogen nutrition of plants and insect invasion. In Nitrogen in Crop Production, ed. R. D. Hauck, pp. 441–460. Madison, WI: American Society of Agronomy. Smith, H. A. & McSorley, R. (2000). Intercropping and pest management: a review of major concepts. American Entomologist, 46, 154–161. Southwood, T. R. E. & Way, M. J. (1970). Ecological background to pest management. In Concepts of Pest Management, eds. R. L. Rabb & F. E. Guthrie, pp. 6–29. Raleigh, NC: North Carolina State University. Tilman, D. (1996). Biodiversity: population versus ecosystem stability. Ecology, 77, 350–363. Vandermeer, J. (1995). The ecological basis of alternative agriculture. Annual Review of Ecology and Systematics, 26, 210–224. Vandermeer, J. & Perefecto, I. (1995). Breakfast of Biodiversity. Oakland, CA: Food First Books.

Chapter 11

Manipulation of arthropod pathogens for IPM Stephen P. Wraight and Ann E. Hajek A great number and diversity of naturally occurring microorganisms are capable of causing disease in insects and other arthropods, and these pathogens have been manipulated for the purpose of pest control for more than 130 years (Lord, 2005). These efforts in applied invertebrate pathology have given rise to the field known today as microbial biological control, or simply microbial control, which is broadly defined as “that part of biological control concerned with controlling insects (or other organisms) by the use of microorganisms” (Onstad et al., 2006). The term biological control does not have a universally accepted definition. The principal disagreement relates to whether or not the term should include control by non-organismal biological factors (e.g. toxic metabolites and other natural products). We believe that the term biological control should be restricted to use of living organisms, and in this context, the simplest definition of biological control is: the use of living organisms to reduce damage by pests to tolerable levels. This definition is similar to those presented by Van Driesche & Bellows (1996), Crump et al. (1999), Eilenberg et al. (2001) and Hajek (2004), all of which limit the definition to use of living organisms. Employing this definition, it follows that biological control agents are living organisms and microbial control agents are living microbes (including viruses, which some

do not consider as living, and nematodes, whose inclusion will be explained below). Most pathogenic microbes in general produce large numbers of propagules (e.g. bacterial and fungal spores and viral occlusion bodies) that function as the infectious units. In most cases, these propagules are resistant to environmental extremes of temperature and moisture and thus capable of surviving for extended periods outside the host. Indeed, the most important function of these propagules is survival in the environment for a period sufficient for successful transport (by myriad mechanisms) to new, susceptible hosts. Observations of this natural mode of action, especially when it culminated in impressive disease outbreaks, stimulated development of methods for mass production and formulation of these propagules for conventional broadcast applications against arthropod pests, an approach essentially equivalent to that employed in the use of synthetic chemical pesticides. In fact, the term microbial control, for much of its history, has been used primarily in reference to use of microbes in this way. These efforts worldwide have produced a number of notable successes, but far more commonly the result has been only partial control (pest population control or suppression insufficient to consistently reduce crop damage to acceptable levels). Consequently, as the result of

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



competition from highly effective chemical pesticides, microbial control products have never comprised more than a very small percentage of the world pesticide market. This situation persists despite the increasing popular demand for environmentally safe pest control methods. Numerous authors have addressed this challenge in recent years (Lacey et al., 2001; Lacey & Kaya, 2007; and see Chapter 13), and there is general agreement that in the future, microbial control agents will make their greatest contributions as components of IPM systems. There are innumerable definitions of IPM; however, we here refer to those definitions that describe IPM systems as pest management systems based on coordinated use of multiple control agents (e.g. chemical toxins and biological control agents) and/or other control factors (e.g. host-plant resistance and cultural practices) to maintain pest populations at economically acceptable levels with minimal impacts on the environment. In such systems, control of any one pest may be achieved through the actions of a single agent or through the combined actions of multiple agents. IPM systems based primarily on biological agents or biologically derived agents (minimizing use of broad-spectrum synthetic chemical pesticides) may be referred to as biologically based (bio-based) or biologically intensive (bio-intensive) IPM systems. In natural systems, populations of insects that are well adapted to the prevailing environmental conditions are regulated primarily by various biotic factors including natural enemies (parasites, predators and pathogens). Virtually all arthropods interact with more than one natural enemy, and development of bio-based IPM systems is often viewed as an attempt to re-establish or replace a natural enemy complex that was lost when an arthropod moved into a new geographic region or new habitat.

11.1 Microorganisms as arthropod biocontrol agents The pathogens that infect invertebrates are as diverse as those infecting vertebrates, and include

all of the major, broadly defined groups of microorganisms: viruses, bacteria, fungi, algae and protozoa. Of these, however, only the viruses, bacteria and fungi (including microsporidia) are currently used for microbial control of arthropod pests. Arthropod-attacking nematodes, though more advanced than single-celled microorganisms, have historically been studied by invertebrate pathologists and are generally viewed as microbial control agents. The nematode species most commonly manipulated for biological control harbor symbiotic bacteria that are released into the body of the invaded host. Once in the host hemocoel, these bacteria act as virulent, lethal pathogens. Nematode reproduction subsequently takes place in the bacterium-colonized host cadaver. Detailed descriptions of the myriad modes of action of these diverse microbial control agents are not provided here, but basics can be found in Hajek (2004) with more detailed descriptions in Boucias & Pendland (1998) and Gaugler (2002). Bacteria, viruses and microsporidia must be ingested to infect their hosts. In contrast, infections by the fungi most commonly used for microbial control are initiated externally, with fungal elements penetrating directly through the host body wall. The infective stages of nematodes also initiate infection by direct penetration, but invasion sites may be external or internal (e.g. through the gut wall). Nematodes also are highly mobile, moving through water or across moist surfaces to locate hosts, especially in soil. These differences in modes of infection influence the types of hosts that can be infected. For example, herbivorous arthropods that feed by piercing plants and sucking sap from the vascular tissues are not susceptible to the common bacterial and viral pathogens applied for biological control. These insects must therefore be targeted with fungal pathogens. As will be discussed, one major difference among pathogens is whether they kill hosts by growing throughout the host (due to infection only) or through production of a toxin (toxinosis). It is difficult to generalize as to the ways in which each group of pathogens has been deployed for microbial biological control. Various pathogens from each group, for example, have been introduced into pest populations under the


objectives of producing both long- and short-term pest control.

11.2 Microbial biological control strategies Approaches for use of microbial control agents are commonly placed in three categories: (1) augmentation biological control, (2) classical biological control and (3) conservation biological control (Hajek, 2004) (Box 11.1). Our principal emphasis in this chapter is with augmentation strategies and, usually, pathogens are applied for inundative control. Pathogens used for strategies other than inundative augmentation must have good potential to recycle in the host population. By recycling, we mean the ability of the pathogens initially infecting hosts to produce inoculum that infects more hosts and for this process to continue.

11.2.1 Types of pathogens used by strategy Augmentation Bacteria, viruses, fungi and nematodes comprise the major groups of pathogens that are massproduced and commercially available for inundative control of arthropod pests. Pathogen-based pesticide products (generally referred to as biopesticides) are applied against many types of arthropod pests, but the major targets are insects in the Orders Lepidoptera, Coleoptera and Diptera. Numbers of products currently produced worldwide for microbial control are presented in Table 11.1 (and see Chapter 9). The microbial control agent that has been used most extensively worldwide is the bacterium Bacillus thuringiensis (Bt), with products worth $US 159.57 million at the end-user level in 2005 (Quinlan & Gill, 2006). This pathogen occurs as many varieties that produce virulent toxins specific to different groups of insects. Despite the considerable commercial success of Bt, its use as a biological control agent

Box 11.1 Definitions for categories of biological control Type of biological control


Augmentation: inundative biological control

The use of living organisms to control pests when control is achieved exclusively by the organisms themselves that have been released. The intentional release of a living organism as a biological control agent with the expectation that it will multiply and control the pest for an extended period, but not that it will do so permanently. The intentional introduction of an exotic biological control agent for permanent establishment and long-term pest control. Modification of the environment or existing practices to protect and enhance specific natural enemies or other organisms to reduce the effect of pests.

Augmentation: inoculative biological control

Classical biological control

Conservation biological control

Source: (Eilenberg et al., 2001; Hajek, 2004).




Table 11.1

Major products available worldwide for microbial control of arthropod pests


Number of species in commercial products




4 (although including many varieties of Bacillus thuringiensis) 12 13

Fungib Nematodes

Number of products with trade namesa 68 238

136 50

Major hosts targeted Lepidoptera, Hymenoptera Lepidoptera, Diptera, Coleoptera Diverse hosts Diverse hosts


In a few cases, products are mixes of different pathogens, and these products are listed for each pathogen. b Including the microsporidian Nosema locustae. Recent molecular studies indicate that the microsporidia, formerly identified as protozoa or protists, are highly reduced fungi. Source: Quinlan & Gill (2006). as in the above definition has declined in recent years due to genetic engineering applications that will be discussed in a later section. Use of the numerous other agents listed in Box 11.1 has grown markedly in the past decade but is still constrained by many factors that will be reviewed herein. Classical biological control Classical biological control has been used only to a limited extent, but has resulted in excellent control in some systems (Hajek et al., 2007). Pathogens that have been most successfully used for classical biological control are species with high epizootic potential that are active in stable habitats such as forests (Hajek et al., 2007). The pathogen most frequently and successfully used for classical biological control has been a nudivirus that attacks palm rhinoceros beetles. An introduced nematode attacking the invasive pine woodwasp Sirex noctilio is considered the most important control agent in IPM programs that also include introductions of parasitoids and thinning of pine stands. A fungus accidentally introduced into North America from Japan is a key agent controlling the gypsy moth (Lymantria dispar), but inundative releases of a nucleopolyhedrovirus and Bt and mating disruption are also part of IPM programs to control this invasive forest defoliator (see Chapter 32).

Conservation biological control This strategy has been investigated only sparingly with microbial control agents, but there have been notable successes that indicate considerable potential in agroecosystems. For example, a program that monitors naturally occurring fungal epizootics of cotton aphid (Aphis gossypii) in the southeastern USA prevents unnecessary insecticide applications that would kill beneficial predators and parasitoids (Steinkraus, 2007).

11.3 Attributes of microbial biocontrol agents Though the microorganisms that have been developed for biological control are extraordinarily diverse, there are a number of general traits that characterize these control agents and greatly influence how they are manipulated for pest management.

11.3.1 Environmental safety Most microbial control agents, like biological control agents in general, are considered environmentally safe or at least environmentally soft. Most have restricted host ranges, and even those species classified as generalists (e.g. the common entomopathogenic fungi Beauveria bassiana and Metarhizium anisopliae) comprise diverse strains


that exhibit more restricted host ranges than the species as a whole. Some species of arthropod pathogens or strains within species produce broadly toxic or mutagenic metabolites (e.g. bacterial or fungal toxins); however, these compounds are not associated with dormant spores or other infectious units upon which most biopesticides are based, and in the environment, they are produced primarily within the infected host (during vegetative growth) and in quantities that are extremely small relative to the amounts of toxic materials typically released into the environment via conventional insecticide spray applications. Because they do not produce toxic residues, biopesticides typically have shorter postapplication re-entry times and pre-harvest intervals than synthetic chemical pesticides, a factor that can translate into a strong economic advantage in labor-intensive fruit, vegetable and ornamental plant production systems.

11.3.2 Speed of action With notable exceptions (toxigenic bacteria and nematodes), most microbial control agents have slow modes of action. Even following a successful inundative application, infected hosts do not succumb to infection for several to many days, or, in some cases, several weeks. During the disease incubation period, infected hosts may continue to damage crops and produce offspring at rates that support pest population growth. Pests infected with microbial control agents may also continue to vector virulent plant pathogens.

11.3.3 Natural epizootic potential and persistence Many microbial control agents are capable of persisting or recycling in the environment (either as resistant spores or at low levels of host infection), and under favorable environmental conditions, some have extraordinary capacities for reproduction and dispersal and can develop rapidly to epizootic levels. This natural epizootic potential is a key factor in use of microbes in classical and inoculative biological control. In these systems, the great capacity of microbes to saturate a pest’s habitat with infectious propagules can produce dramatically sudden pest population crashes (despite

relatively slow action against the individual members of the population).

11.3.4 Host age-dependent susceptibility Though there are numerous exceptions, most arthropod pathogens are more virulent against the immature (larval and nymphal) stages than the adults of their hosts. Within larval or nymphal instars, susceptibility also tends to decrease with increasing age. High virulence against immatures is an important trait of many biological control agents, because it confers the potential to compensate for slow action by preventing the pest from reaching larger stages that cause greater damage or from reaching reproductive maturity. Immatures of many pests also aggregate (e.g. at favored feeding sites on a host plant), making them vulnerable to pathogen epizootics. On the other hand, larvae of many important pests lead solitary lives in cryptic habitats, reducing susceptibility to pathogens capable of causing epizootics and to inundative or inoculative applications of pathogens. It should also be noted that few pathogens used for microbial control are highly virulent against the egg stage of their hosts.

11.3.5 Environmental sensitivity Efficacy of many microbial control agents is highly dependent upon environmental conditions. Conditions of low or high temperature, low moisture and especially high insolation can severely limit both the initial and residual activity of microorganisms applied in inundation and inoculation biological control programs. Adverse abiotic conditions can also greatly diminish the natural epizootic potential of microbes released for inoculation or classical biological control.

11.3.6 Host contact With few exceptions, lethal actions of microbial control agents are initiated only after infectious propagules (or infectious stages of nematodes) come into contact with the host. In most cases, microbial propagules are incapable of directed movement and therefore unable to actively search for susceptible hosts or avoid unfavorable environmental conditions; this trait is often cited as the reason many pathogenic microbes produce great numbers of propagules. Lack of mobility is




especially significant when suitable hosts are present only in cryptic habitats or at low densities. In such cases, novel formulations (e.g. with baits or other attractants) or highly efficient, precisely targeted application methods may be required to achieve efficacy. This contrasts sharply with chemical control agents exhibiting vapor, translaminar, or systemic activity, for which less efficient application methods may suffice.

11.3.7 Mass production Many microbial control agents can be massproduced in vitro on industrial fermentation scales, and thus their populations can be manipulated more easily, rapidly and economically than those of most macrobiological control agents (i.e. insect parasitoids and predators). Inundation biological control is a major strategy in microbial control, whereas few parasitoids and predators are used in this way. On the other hand, some microbial control agents (especially obligate pathogens with high host specificity) are difficult or impossible to mass-produce and formulate, and some agents formulated as biopesticides have limited shelf-lives. Use of these agents for pest control may be limited to inoculative releases, conservation, or classical biological control approaches.

11.3.8 Product registration Though microbial control agents in general have excellent environmental safety records, they are generally considered to pose greater health risks to humans and other non-target organisms than macrobiological control agents. Some arthropod pathogenic microbes are allergenic, and, as previously indicated, some have the potential to produce toxic metabolites. Mass production systems for microbial control agents also may be accidentally contaminated with other microbes that pose unknown risks. In many countries, therefore, microbial control agents mass-produced and formulated as biopesticides are required to undergo rigorous registration and quality control processes. Consequently, commercial development of these agents may be more costly than development of parasites or predators. As an exception, nematodes, once permitted into a country and cleared for release, are usually exempt from registration requirements.

11.3.9 Response to dose Infectious pathogens exhibit dose–response regression lines with low slopes. The infectious units of pathogenic microbes generally act independently or largely independently in establishing lethal infections. One of the most significant effects of this independent action is to constrain the slope (regression coefficient) of the dose– response regression. Slopes of log dose-probit response regression lines of arthropod pathogens rarely exceed 2 and are often substantially 85% efficacy, as compared to typical chemical standard (standard used should be explained), OR r >90% achievement of indicated response expectation indicated on advertising collateral (6) Testimonials from peer (6) Signed testimonials from each of the groups, attesting to quality following relating to quality, consistency claims, specific efficacy and efficacy: r at least five independent end-users; claims and overall also include proof of purchase satisfaction with product use r at least three university co-operators r at least two trade channel representatives; also include proof of purchase (7) Achievement of field trial (7) Documentation of scientifically valid variability within +/−25% efficacy trials and claims representing of collateral or advertising statistical significance relating to claims promotional claims (Note: no penalty for exceeding claims, but excessive variability through over formulation is not desirable) (8) Product complaint history (8) Documentation showing complaints demonstrating product’s represent less than 5% of product sold commercial use meeting (volume basis) over the previous grower expectations two-year period


Box 13.2 Summary of results from trial comparing biopesticide with conventional chemical insecticide Treatment


Conventional insecticide alone 83% Conventional insecticide – biopesticide rotation 79% Control 0% “There was no advantage to rotating the conventional treatment with a biopesticide.” However, viewing the trial with a “glass half-full” the trial would be reported as follows: The level of control was maintained while gaining the following: • Resistance management • Re-entry period − flexibility with labor • Reduced time to harvest − multiple harvest • Residue management

Biopesticide companies do not have the dollars of the big companies to conduct trials. Marketable yield is the most important measure of performance; however, pest/disease level is usually the measure in performance trials. Biopesticides often increase marketable yield (and results often can be as good as or better than with chemicals) but trials may have more diseases/pests than in trials with conventional chemicals. Chemicals fail, yet the perception is they always work. Excitement often accompanies the introduction of new conventional chemistry, but often there is less enthusiasm over a new biopesticide even when has good efficacy.

13.10 Conclusions By combining performance and safety, biopesticides offer benefits not generally realized with conventional pesticides. These benefits include efficacy while providing customers the flexibility of:

r r r r

minimum application restrictions, residue management, resistance management, and human and environmental safety.

Consumers are driving growth of organic food and food produced with reduced-risk pesticides

and fewer pesticides. Due to government regulations around the world, older, more toxic chemical pesticides are being removed from the market and there is increased emphasis on worker safety and pesticide exposure and attention to pesticide contamination of air and water. In addition, the global production and shipment of food products has resulted in an emphasis on maximum residue levels on food at time of shipment and entry into the receiving country. These trends are driving the global growth of biopesticides at rates much faster than the mature, slow growth chemical pesticide market. Despite these trends and continued fast biopesticide growth, perception of biopesticide products and lack of awareness remain the largest barriers to increased adoption. Key influencers (e.g. university extension specialists and PCAs) and gatekeepers (e.g. distributors) may have decided opinions about the efficacy of biopesticides, or lack thereof, and end-users (growers, consumers and superintendents) generally may not understand their benefits and value, how they work and how to use them. This provides a good opportunity to policy makers and the industry to finds ways to educate and increase awareness. For example, the IR-4 program, funded by the EPA, has developed a searchable biopesticide database (Environment Protection Agency, 2007). For the first time, an end-user or key influencer can find information about biopesticides all in one place.




It is a good first step. It is time to move from listing and discussing the barriers to adoption, which are now well known and focus on breaking down these barriers. These include:

r Increased federal funding through the IR-4 pror r r


gram for biopesticide efficacy trials and demonstration programs. BPIA standards and seal of approval for biopesticide products. Increased applied research at land-grant universities integrating biopesticides with chemicals and with other biopesticides. Increased effort by biopesticide companies to educate their customers and specifically position their products in the marketplace and selling through experienced channel partners. Increased education of end-users by biopesticide companies, extension specialists and other key influencers (USDA, EPA, California Department of Pesticide Regulation).

References Agrow (2006a). Global agrichemical market flat in 2005. Agrow: World Crop Protection News, February 24, 2006, 490, 15. Agrow (2006b). US agrochemical sales up 2% in 2005. Agrow: World Crop Protection News, October 6, 2006, 505, 14. Beyond Pesticides (2008). Home page. Available at Bounds, G. & Brat, I. (2006). Turf Wars: armed with chicken manure, organic lawn product makers are battling for your backyard. Wall Street Journal, April 15, 2006, P1. Available at article/SB114505266809026471.html. Braverman, M. (2005). Why use biopesticides in IPM programs? In Proceedings of 2005 National IPM Symposium. Available at ipmsymposium/sessions/51_Braverman.pdf.

Cuddeford, V. (ed.) (2006). Surveys gauge attitude of Canadian public. Biocontrol Files: Canada’s Bulletin on Ecological Pest Management, January 2006, 5, 8. Available at Donaldson, M. (2003). Results of a customer survey on attitudes towards biopesticides. In Proceedings of 2003 National IPM Symposium. Available at symposium/getsession.cfm?sessionID=47. Environmental Protection Agency (2007). What Are Biopesticides? Washington, DC: US Environmental Protection Agency. Available at biopesticides/whatarebiopesticides.htm. Fredonia Group (2006). Focus on lawn and garden consumables. Fredonia Focus Reports, Consumer Goods, August 1, 2006. Available at business/ama/home.htm. Cleveland, OH: The Freedonia Group. Guillon, M. (2003). President’s Address. Available at National Gardening Association (2002). 2002 National Gardening Survey. South Burlington, VT: National Gardening Association. National Gardening Association (2004). Environmental Lawn and Garden Survey. South Burlington, VT: National Gardening Association and Organic Gardening Magazine. Quinlan, R. & Gill, A. (2006). The World Market for Microbial Biopesticides, Overview Volume. Wallingford, UK: CPL Business Consultants. US Department of Agriculture (2003). Symposium: Barriers to the adoption of biocontrol agents and biological pesticide. In Proceedings of the 4th National IPM Symposium/Workshop. Washington, DC: US Department of Agriculture. Available at symposium/getsession.cfm?sessionID=47. Venkataraman, N. S., Parija, T. K., Panneerselvam, D., Govindanayagi, P. & Geetha, K. (2006). The New Biopesticide Market. Wellesley, MA: Business Communications Company Inc. (BCC) Research Corporation. Available at asp.

Chapter 14

Integrating pesticides with biotic and biological control for arthropod pest management Richard A. Weinzierl Insecticides and acaricides are used commonly for pest suppression in agriculture, forestry and public health. The National Research Council (2000) and others have documented their value in protecting crops, livestock and humans from injury by insects and other arthropods. However, adverse effects of pesticide use can include the killing of non-target organisms, contamination of water supplies and persistence of unwanted residues on foods and animal feed. The chronic health effects of even extremely low concentrations of pesticides in food and water remain under debate today just as they were when the National Research Council (1993) reported on pesticides in children’s diets in the early 1990s. Resistance to one or more pesticides has evolved in populations of over 500 insect and mite species (Clark & Yamaguchi, 2002), rendering many of those pesticides ineffective against the resistant populations. For these and other reasons, supplementing or replacing pesticides with non-chemical control tactics, including biological control, is a goal in many crop and livestock production systems. As biological control programs that rely on parasites, predators and pathogens are developed, they rarely are so robust as to provide all the pest suppression needed in an entire cropping system for the lifespan of the crop. Even if such robust biological control is attainable over time

in specific crops or situations, initial efforts that target only one or a few pests often leave a need for pesticides to manage other species if their populations exceed economic thresholds. Insecticides used under such circumstances may be (and often are) toxic to organisms used for biological control (Newsom et al., 1976; Hull & Beers, 1985; Croft, 1990), and the effectiveness of the biocontrol program is thereby threatened. For this reason, effective integration of biological and chemical controls is essential. Previous works by Stern et al. (1959), Stern & van den Bosch (1959), Smith & Hagen (1959), Gonzalez & Wilson (1982) and Hoy (1989) illustrate that the need for such integration has long been a key issue. Integrating biological and chemical control is, after all, the foundational idea in the “integrated control” concept proposed by Stern et al. (1959) now known as IPM. The integration of chemical and biological control tactics can take many forms, depending on the characteristics of the biological control agent and its host, the nature of the cropping system or habitat, the persistence and toxicity of the pesticides to be used and the methods by which pesticides are applied. Approaches to integration are presented and discussed in this chapter under four headings: (1) reducing pesticide use; (2) using selective pesticides and applying pesticides selectively; (3) modifying biological control agents to survive

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



pesticide applications; and (4) combining biological control agents and pesticides for increased effectiveness. The chapter closes with a brief discussion of the current status of efforts to integrate the use of chemical pesticides and natural enemies in arthropod pest management. In this chapter, the term biotic control is used to describe the naturally occurring effects of predators, parasites and pathogens; biological control is used to describe human-managed efforts to preserve or use natural enemies. The terms blur a bit when IPM includes well-planned efforts to conserve natural enemies by specific habitat management practices, pesticide selection and other actions. For purposes of illustrating these concepts, the scope of the discussion is limited to crops and livestock.

14.1 Reducing pesticide use Reducing the use of pesticides is an obvious way to reduce their negative impacts on biotic and biological control organisms. Clearly, pesticides should not be applied unnecessarily, and ending unneeded, ineffective and unprofitable uses in any crop or setting would save money, reduce environmental risks and allow survival of natural enemies. It would be incorrect, however, to assume that all pesticide applications are unnecessary and could be ended without dramatic increases in losses to pests (National Research Council, 2000). What steps, then, are realistic for reducing the use of pesticides and allowing greater success for biological control? The main categories of actions are: (1) designing production systems that minimize the need for pesticides; (2) identifying natural enemies as beneficial species; (3) developing and using thresholds and models that accurately reflect control needs and account for the impacts of natural enemies; and (4) delivering educational programs to farmers, landscapers, greenhouse managers, pest management consultants and the public.

14.1.1 Production systems that rely less on pesticides Approaches to reducing pesticide use incorporate such practices as planting resistant cultivars, practicing effective crop rotations, altering time

of planting to avoid pest presence, using physical barriers to exclude pests and maintaining habitats that favor natural enemy survival. All of these practices reduce the need for pesticides that might also kill biotic or biological control agents. For example, where application of broadspectrum insecticides for control of onion thrips (Thrips tabaci) in cabbage would interfere with biological control of lepidopterans such as cabbage looper (Trichoplusia ni) and diamondback moth (Plutella xylostella), choosing thrips-resistant cultivars (Egel et al., 2007) might avoid the need for insecticide use and unwanted side effects. Similarly, planting sorghum cultivars resistant to sorghum midge (Stenodiplosis sorghicola) and/or greenbug (Schizaphis graminum) reduces the likelihood that insecticides will be needed for the control of these pests (Teetes, 1994). Soil applications of insecticides to protect cucurbits (cucumbers, squash, melons and pumpkins) from seedcorn maggot (Delia platura) likely reduce populations of beneficial predaceous carabid and staphylinid beetles (Hassan, 1969; Brust et al., 1985; Curtis & Horne, 1995), but where these crops are directseeded into cool, wet soils, seedcorn maggot control often is necessary. Using transplants instead of direct-seeding reduces the risk of damage by seedcorn maggot and usually avoids the need for insecticidal control. Delaying planting of cucurbits would likewise escape injury by seedcorn maggot. Floating row covers and pest-proof screening exclude a wide range of insect pests in outdoor and greenhouse production of horticultural crops, reducing the need to apply pesticides that might kill biological control agents. Because many natural enemies benefit from stable habitats, maintaining ground covers, standing crops and crop residues (instead of practicing clean tillage) generally results in greater numbers of predators, parasites and pathogens. Additionally, maintaining plants that provide nectar and pollen for adults of parasitic wasps increases their retention and survival (Landis et al., 2000).

14.1.2 Identifying natural enemies Recognizing natural enemies is an obvious basic requirement for minimizing pesticide use. Although certain predators, e.g. lady beetles and praying mantids, may be widely recognizable,


many farmers and gardeners do not recognize adults of parasitic insects (Hymenoptera and Diptera). Likewise, larvae of lacewings and lady beetles and adults and immatures of carabid and staphylinid beetles, syrphid flies and predaceous Hemiptera are common but often not identified by farmers and gardeners. If a failure to identify them as beneficial organisms leads to killing them with insecticides, many other biological control agents often are killed as well.

14.1.3 Improving and using pest thresholds and predictive models Refinement of economic thresholds and development of models to predict changes in pest densities based on field-level population sampling may be among the most important steps for better integration of biological and chemical control in agriculture. Insect management guidelines that include pest density thresholds for insecticide applications often mention natural enemies and suggest that pesticide application may not be needed if they are present (Steffey & Gray, 2007). However, few such recommendations available for growers or consultants give specifics on the number of natural enemies needed to provide enough biotic or biological control to prevent crop losses that would exceed the cost of a pesticide application. Establishing and communicating clear guidelines on the numbers of natural enemies needed to control pest populations of varying densities would allow a grower or consultant to reject a decision that the pest population exceeds the economic injury level and decide instead to allow biotic or biological control to occur (avoiding unneeded expense and the nontarget impacts of pesticide application). Examples of information specific enough to really guide decision making include assessments of the numbers of eggs of common green lacewing (Chrysoperla carnea) needed if subsequent larvae are to prevent economic losses to Heliothis spp. in cotton (Ridgway & Jones, 1969), the impacts of specific numbers of the parasitoids Bathyplectes cucurlionis and Tetrastichus incertus on alfalfa weevil (Hypera postica) (Davis, 1974 and Horn, 1971, respectively), necessary ratios of common green lacewing for suppression of apple aphid (Aphis pomi) in apples (Niemczyk et al., 1974), and the number of con-

vergent lady beetles (Hippodamia convergens) per sweep necessary for adequate control of pea aphid (Acyrthosiphon pisum) in alfalfa (Hagen & McMurtry, 1979). Tamaki et al. (1974) used the term “predator power” to describe the idea that different natural enemies kill different numbers of aphids and therefore have greater or lesser pest management impact. Polyphagous hemipteran predators such as nabids, anthocorids and geocorids were given a predator power rating of 1; immature coccinellids, syrphids and chrysopids were rated 4 and large adult coccinellids were rated 8. Combined predator power ratings for populations sampled in an individual field were calculated to predict aphid suppression (Tamaki et al., 1974). Tamaki & Long (1978) further developed a predator efficacy model that included temperatures and functional responses to predict impacts of inundative releases. Tamaki and colleagues focused on the impacts of inundative releases and did not suggest modifying pest thresholds or decisions on pesticide applications based on natural enemy densities. However, Naranjo & Hagler (1998) assessed impacts of heteropteran predators for just such a purpose. Adjusting thresholds for aphid control based on parasitism (Fernandez et al., 1998; Giles et al., 2003) or infection by fungal pathogens (Hollingsworth et al., 1995; Conway et al., 2006) has been proposed for cereal crops and cotton. Hoffman et al. (1990, 1991) proposed incorporating impacts of Trichogramma spp. on Helicoverpa (= Heliothis) zea to modify thresholds for egg counts in processing tomatoes. Gonzalez & Wilson (1982) took a season-long approach to this idea and suggested that economic thresholds should consider the food webs that exist in cropping systems, particularly cotton. Nyrop & van der Werf (1994) summarized methods to monitor natural enemies for the prediction and assessment of biological control. Brown (1997) outlined two categories of models designed to incorporate the impacts of natural enemies in estimates of economic thresholds, one for generalist natural enemies with population densities not tied closely to the dynamics of a given pest and a second for specialist natural enemies whose population densities are coupled tightly with the pest’s population




dynamics. Musser et al. (2006) suggested that natural enemy densities and their projected impacts should be incorporated into thresholds and that thresholds should differ for individual pesticides according to the negative impacts of those pesticides on natural enemies. This approach incorporates and expands upon ideas expressed well over a decade ago by Kovach et al. (1992) and Higley & Wintersteen (1992). Despite the research presented in professional journals, relatively few pest management recommendations that reach growers incorporate assessments of biotic control. Weires (1980), in a fact sheet written for apple growers, summarized recommendations for control of European red mite (Panonychus ulmi) based on numbers of European red mites and their predators on a perleaf basis. Possible actions recommended for mite management included applying a miticide, waiting to resample, or concluding that biotic control is likely. If growers or pest management consultants are to incorporate counts of natural enemies into their decisions on the need for insecticide application (and thereby integrate biotic and biological control with pesticides), specific guidelines such as these are essential. For example, for growers to weigh the impacts of lady beetle predation on soybean aphid (Aphis glycines) and determine whether biotic control will suppress aphid densities enough to prevent economic losses, detailed recommendations based on predator– prey population dynamics are needed for a range of population densities and environmental conditions.

14.1.4 Educating growers and consumers Providing growers and consumers with the information needed to make educated decisions is a constant need in guiding insecticide use. Numerous guides to IPM, natural enemy identification and the practice of biological control provide recommendations on when pesticides are needed (and which ones are most effective or appropriate) and when alternatives – including no control at all – may be better choices. Nonetheless, educational programs that discourage unnecessary pesticide applications face strong competition when marketing efforts of pesticide companies strive for maximum sales. Marketing efforts that use biased

presentations of research data remain as much of a concern as they were when Turpin & York (1981) wrote “Insect management and the pesticide syndrome.” For example, in Illinois, despite conservative thresholds for soybean aphid control (i.e. 250 aphids per plant) (Ragsdale et al., 2007), soybean growers tell of pesticide sales representatives who recommend insecticide application at much lower densities, using the rationale that a few aphids today will increase to exceed the threshold soon. They ignore the impacts of biotic control that are conservatively incorporated in the recommended threshold and they urge that “the sooner the better” is the rule to follow for insecticide application. It is indeed an understatement that education on the benefits of not spraying is an ongoing need.

14.2 Using selective pesticides and applying pesticides selectively Where insecticides must be used in pest management but their integration with biotic and biological control also is a goal, the use of selective insecticides or selective application methods commonly is advised (Newsom et al., 1976; Weinzierl & Henn, 1991). More than a half-century ago, Ripper et al. (1951) defined selective insecticides as “insecticides toxic to the pests but not toxic to the beneficial insects.” They noted that beneficial insects that survive spray treatments (with selective insecticides) exert: (1) an immediate effect when they “mop up the surviving individuals of the pest population and thus increase the apparent effectiveness of the insecticide application,” and (2) a delayed effect when “(they) or their progeny act on the surviving pest population and prevent the start of the rapid build-up of the pest population which is so often observed when conventional insecticides are used.” They considered selective insecticides to include the systemic organophosphate schraden, as well as bait formulations of DDT encapsulated in hemicelluloses that were poisonous only to phytophagous insects that consumed the bait. Similarly, Stern & van den Bosch (1959) and Smith & Hagen (1959) considered the organophosphate systox to be selective because


when used as a systemic its toxicity to key natural enemies was low. More recently, the concept of selectivity has been discussed under two broad headings: (1) physiological selectivity, in which specific pests and natural enemies differ in susceptibility to individual broad-spectrum toxins or in which the mode of action of the pesticide is specific to certain groups of insects; and (2) ecological selectivity, which depends largely on either spatial or temporal separation of the insecticide’s effects and the occurrence of key natural enemies (Theiling & Croft, 1988; Croft, 1990).

14.2.1 Physiological selectivity The physiological selectivity of pesticides is discussed in detail by Theiling & Croft (1988) and Croft (1990). Their SELECTV database summarizes results from studies of more than 12 000 insecticide–natural enemy combinations. The authors calculated selectivity ratios to reflect the relative toxicity (based on LD50 s, LC50 s and similar measures) of selected pesticides to natural enemies in comparison with their hosts or prey (the target pests). Hundreds of additional reports on the toxicity of individual insecticides to specific natural enemies have been reported since the SELECTV database was first published. Representative examples of studies included in the database and those that have been published more recently include works by Hassan & Oomen (1985), Hagley & Laing (1989), Boyd & Boethel (1998), Medina et al. (2003) and Tillman & Mullinix (2004). When data from multiple trials examining multiple pesticides are analyzed, assessments of the relative toxicity of available pesticides can guide growers’ selections of the least toxic broadspectrum insecticides where such compounds must be used. For example, choosing insecticides for control of direct pests of apples, including codling moth (Cydia pomonella), plum curculio (Conotrachelus nenuphar) and apple maggot (Rhagoletis pomonella), has for many years included consideration of the toxicity of those insecticides to predaceous mites that control European red mite. The organophoshates azinphosmethyl and phosmet became standards in apple pest management because predator species developed resistance to them, rendering them less disruptive to biotic control than most other

organophosphates, carbamates, pyrethroids and neonicotinoids (Croft & Meyer, 1973; Rytter & Travis, 2006). Hoy (1995) suggested that information on the toxicity of new pesticides to natural enemies important in specific cropping systems be required for US Environmental Protection Agency (EPA) registration of those pesticides. This information could be included in a database available to pesticide users, crop consultants and IPM advisers. One problem for such a database is that selectivity of broad-spectrum insecticides and miticides to specific natural enemies may differ significantly over time and among geographical regions as these organisms respond to selection pressure (Croft & Meyer, 1973; Croft, 1990). In more recent years, physiological selectivity based on taxon-specific modes of action has provided greater options for integrating pesticides and biological control. Examples of insecticides that are selective based on mode of action include microbial insecticides and synthetic compounds. Among the microbials (some of which also may be considered as biological control agents themselves) are insecticides derived from subspecies of the soil-borne bacterium Bacillus thuringiensis (Bt); different products are toxic only to larvae of Lepidoptera, certain lower Diptera, or a few species of Coleoptera (Maagd et al., 2001). Other microbial insecticides contain taxon-specific viruses, fungi or microsporidia. Selective synthetic insecticides include compounds primarily toxic to Lepidoptera (tebufenozide, methoxyfenozide, diflubenzuron, novaluron), Hemiptera/Homoptera (buprofezin, kinoprene and pyriproxyfen) or Diptera (cyromazine) (Medina et al., 2003; Stark et al., 2004; Insecticide Resistance Action Committee, 2006). Cloyd & Dickinson’s (2006) summary of the toxicity of buprofezin, pyriproxyfen, flonicamid, acetamiprid, dinotefuron and clothianidin to a parasite and a predator of citrus mealybug (Planococcus citri) illustrates the value of the selectivity of taxon-specific pesticides for a greenhouse system. Selective miticides that are low in toxicity to predaceous mites and certain other predators in orchards include chlofentezine, acequinocyl and hexythiazox. Miticides that are somewhat more toxic to predaceous mites but are not toxic to predaceous insects include bifenazate, spirodiclofen and fenpyroximate (Rytter & Travis,




2006). Additionally, pheromone- or kairomonebased management using mass trapping, mating disruption, or attract-and-kill concepts offers physiological selectivity and little or no direct mortality to natural enemies. It is important to note that insecticides considered to be selective, including lepidopteran or hemipteran/homopteran growth regulators, may be toxic to natural enemies in taxa other than those targeted by the insecticide’s primary mode of action. For example, Rothwangi et al., (2004) found that growth regulators thought to narrowly target hemipterans/homopterans were toxic to the mealybug parasite Leptomastix dactylopii. Conversely, even broad-spectrum insecticides and miticides may be considered at least somewhat selective within a biological control program that relies on pathogens as the organisms employed by humans (Barbara & Buss, 2005; Ericsson et al., 2007). Finally, transgenic crops that produce Bt toxins might be deemed selective pesticides. While the toxins they produce directly kill only certain taxa, the extremely high level of control they provide, often in a high percentage of crop area in a given region, has obvious ecological impacts on natural enemies of target pests. Their sustainable use as selective insecticides is dependent on steps that provide ecological selectivity as discussed later in this chapter. Assessing the toxicity of specific pesticides against specific natural enemies in laboratory tests has been the most common approach to identifying physiological selectivity (Theiling & Croft, 1988; Croft, 1990), but Stark & Banks (2003) suggest that data gained from such bioassays need to be interpreted carefully. They propose that population-level effects of pesticides must be measured to understand the impacts of physiological selectivity. Recent examples of studies that have assessed physiological selectivity (perhaps in combination with ecological selectivity) at the population level include those by Bernard et al. (2004), Furlong et al. (2004), Wilkinson et al. (2004) and Desneux et al. (2005).

14.2.2 Ecological selectivity As noted previously, ecological selectivity results from the separation of a pesticide’s effects from

the occurrence of susceptible natural enemies. Either time or space may be the separating factor. Where broad-spectrum pesticides are used, integrating biological and chemical control may depend on timing of pesticide applications in comparison with natural enemy occurrence, rates of application and persistence of the pesticides applied. Clausen (1956) noted that with the introduction of DDT and other persistent insecticides, the impacts of pesticide applications on natural enemies were dramatically greater than they had been when less persistent botanical insecticides were the primary or only chemicals used. Although pesticides currently in use are far less persistent than DDT and other organochlorines that Clausen referenced, even moderate persistence means that if time is to provide ecological selectivity, the broad-spectrum insecticide usually must be applied after natural enemies have exerted the majority of their potential benefit in a given crop season. For example in the midwestern USA, conserving natural enemies of diamondback moth while still using insecticides for lepidopteran control is a goal in production of cabbage, broccoli and related crops. To accomplish this goal, growers are urged to use microbial insecticides containing Bt when infestations of lepidopteran larvae exceed thresholds prior to heading (Weinzierl & Cloyd, 2007). This approach provides adequate control before heading and allows natural enemies of diamondback moth, cabbage looper and imported cabbageworm (Pieris rapae) to survive and attack larvae not killed by Bt applications. Pyrethroids are used as needed after heading, especially for control of cabbage looper. Pyrethroids are the most effective insecticides against this pest, especially if middle or late instars are present, and high levels of control are needed at this time to prevent damage to and contamination of heads. Withholding pyrethroid use earlier in crop development allows biotic control agents to suppress diamondback moth, resulting in temporal integration of biotic and chemical control. The scale of spatial separation required to provide ecological selectivity for broad-spectrum pesticides may range from millimeters to decimeters. The following examples all represent selective


applications of pesticides in ways that allow survival of natural enemies.

r Systemic seed treatments and in-furrow applications of systemic insecticides kill insects in direct contact with the treated seed or soil and those that feed on young plants, but natural enemies outside the treated furrow or on the surface of the treated crop are not directly affected (though consumption of prey or parasitism of hosts feeding on plant tissue may result in indirect poisoning). Banded applications of insecticides (applied to only a narrow strip of soil surrounding the row) used to protect against root-feeding insects such as Diabrotica spp. allow survival of carabids and other soilinhabiting predators outside the treated area. r Spot treatment of two-spotted spider mite (Tetranychus urticae) infestations in soybeans – usually treating field edges where mites build up first as they move into fields from surrounding vegetation – instead of treating entire fields allows survival of lady beetles and other predators and parasites that provide biotic control of soybean aphid. r Alternate-row applications of insecticides in apples and other tree fruit and nut crops (where sprays are applied from only one side of each row of trees, resulting in less than complete coverage) allow survival of predators of European red mites on the leeward side of trees, and redistribution of these predators within trees provides some level of continuous survival and predation. Hull & Beers (1985) described additional practices for achieving ecological selectivity in orchard systems. r Where pteromalid parasites (Hymenoptera: Pteromalidae) are used in biological control of housefly (Musca domestica) or stablefly (Stomoxys calcitrans), insecticides can be used as residual sprays to walls of livestock buildings to kill adult flies without interfering significantly with parasite survival (Jones & Weinzierl, 1997). Although adult flies are poisoned when they rest on treated surfaces, these parasites attack fly pupae in manure, and their hostfinding flights do not bring them into contact with spray residues. Insecticides applied as manure sprays or as aerosols would cause

more mortality to these parasites, but they are separated in space from the effects of surface residual sprays.

14.2.3 Refuges Finally, a unique form of ecological selectivity resulting from spatial separation of natural enemies from pesticides occurs when refuges are left untreated to allow pest survival. Leaving refuges of untreated crops is recommended almost exclusively for the purpose of managing pesticide resistance, and the only widespread use of the practice to date has been in managing (slowing the development of) pest resistance to Bt toxins in transgenic plants (Ostlie et al., 1997). Transgenic crops are regulated in the USA as pesticides and the planting of non-Bt refuges is required as a condition of their EPA registration. In general, neither the lepidopteran-specific Bt toxins (e.g. Cry1Ab and Cry1F in corn and Cry1Ac in cotton) nor the corn-produced Bt toxins that kill corn rootworm larvae (Cry3Bb1, Cry34Ab1 and Cry35Ab1) have been found to be directly toxic to predators or parasites of key pests in controlled laboratory studies or limited-duration field surveys (Shelton et al., 2002; Al-Deeb & Wilde, 2003; Al-Deeb et al., 2003). However, Andow & Hilbeck (2004), L¨ ovei & Arpaia (2005) and Naranjo et al. (2005) argue that the methods used in many assessments of impacts of transgenic crops on natural enemies have been inadequate and that long-term ecological impacts are possible, if not likely. There is no argument that transgenic crops offer neartotal control of several target pests and that hosts and prey for specialist natural enemies of those pests are absent in “treated” fields (fields planted exclusively to a Bt crop). Refuges therefore serve not only in resistance management as a necessary habitat for pests that are susceptible to Bt toxins, but also as habitat where biotic and biological control agents might survive. Integrating chemical control in the form of transgenic crops with biotic and biological control is dependent on the use of refuges. Considered in total, physiological selectivity and ecological selectivity can be exploited to varying degrees in different crops or settings to gain the benefits of pesticides and still allow survival of biotic control agents and maintenance of a




meaningful biological control program. Given the opportunity to use selective chemicals or selective approaches to their application, one might ask why non-selective chemicals and application practices are used so widely despite their negative impacts on biological control. Although part of the answer lies in the need to better educate growers and other pesticide users about assessing the need for control, the benefits of biological control and the negative impacts of pesticide use, other factors also lead to broad-spectrum pesticide use. Where more than one pest organism may be targeted by a single control method, such as the application of a broad-spectrum insecticide to control codling moth, apple maggot and white apple leafhopper (Typhlocyba pomaria) in apples, applying three separate selective insecticides, all of which might be less toxic to many natural enemies, may represent three times the dollar cost. If they are present at levels that warrant control, killing several pests with a single insecticide application appeals even to conscientious growers.

14.3 Modifying biological control agents to survive pesticide applications Many insecticides are more toxic to parasites and predators than they are to pests attacked by these natural enemies (Croft, 1990). However, resistance to pesticides can develop in natural enemy populations just as it does in pests. Pielou & Glasser (1952) reported on selection of DDT resistance over 50 years ago in the parasitic wasp Macrocentrus ancylivorus. Resistance can evolve in natural enemies in field conditions, or laboratory selection may be used to produce resistant populations for release in the field. Increased resistance in natural enemies renders broad-spectrum insecticides at least somewhat selective and therefore allows their use in conjunction with biotic or biological control. Field selection resulting from exposure to repeated insecticide or miticide applications has led to evolution of resistance to one or more pesticides in several predaceous or parasitic arthropods, including Metaseiulus occidentalis,

Typhlodromus pyri, Neoseiulus (= Amblyseius) fallacis, Coleomegilla maculata, Bracon mellitor and M. ancylivorus (Croft & Brown, 1975). Croft & Meyer (1973) observed that azinphosmethyl resistance in N. fallacis increased 300-fold in a Michigan apple orchard treated five to seven times per year over a 4-year period. Field selection of resistance in natural enemies is a major cause of differing observations of physiological selectivity for specific pest and pesticide combinations over time or among locations, as represented in the data summarized by Croft (1990). Laboratory selection programs for increasing pesticide resistance in natural enemies have been used most extensively to develop resistant strains of predaceous mites that are important in the suppression of tetranychid mites in deciduous tree fruits and nuts (Hoy, 1985). Selection programs comprised of both field and laboratory components have produced strains with resistance to multiple pesticides (Hoy, 1989; Whitten & Hoy, 1999). Although initial efforts centered on phytoseiid mites, laboratory selection has also been used to develop resistant strains of insects such as C. carnea (Grafton-Cardwell & Hoy, 1986), Aphytis melinus (Rosenheim & Hoy, 1988) and Trioxis pallidus (Hoy & Cave, 1988). In addition to standard selection practices for development of pesticide-resistant natural enemies, molecular genetic techniques now make it possible to transfer resistance genes among species. Li & Hoy (1996) and Hoy (2000) used maternal microinjection to transform the predaceous mite M. occidentalis. Currently, rules for the release and use of transgenic natural enemies have not been established, and none are in use (Hoy, 2006).

14.4 Combining biological controls and pesticides for increased effectiveness Many natural enemies of arthropods are other arthropods – predators and parasites – and many of those are killed when broad-spectrum insecticides are applied (with exceptions as noted above in the discussion of selectivity). However, many insect pathogens – viruses, bacteria, fungi,


microsporidia and nematodes – are not harmed by the use of most insecticides or acaricides. Where pathogens are used in classical biological control or in augmentation as microbial insecticides, conventional insecticides may be applied as well if needed, and in many instances there is no direct mortality to the pathogens. For example, codling moth granulosis virus may be used in orchards where conventional insecticides also are applied, and it infects and kills codling moth larvae not controlled by the conventional treatments (although virus buildup may be limited by low host densities). Insect-pathogenic fungi (e.g. Beauveria bassiana) or nematodes (e.g. Steinernema spp. and Heterorhabditis spp.) may be used along with many soil insecticides to increase the range of pest species controlled. In some instances, insecticides and pathogens may act synergistically. For example, Quintella & McCoy (1998) found that imidacloprid and two entomopathogenic fungi acted synergistically against the root weevil Diaprepes abbreviatus. Barbara & Buss (2005) found that using insecticides and parasitic nematodes together gave increased control of mole crickets (Scapteriscus spp.). Brinkman & Gardner (2001), Furlong & Groden (2001) and Ericsson et al. (2007) reported similar findings of insecticide and pathogen synergism. Combining pathogens and chemical pesticides does not always result in additive or synergistic effects. For example, James & Elzen (2001) reported antagonism between B. bassiana and imidacloprid when combined for control of silverleaf whitefly (Bemisia argentifolii).

14.5 Conclusions The integration of biotic and biological controls with chemical pesticides has presented challenges for several decades, at least since the rapid increase in use of synthetic insecticides that began in the 1950s. Methods to accomplish this integration, as outlined in this chapter – reducing pesticide use, increasing the selectivity of pesticide use, modifying natural enemies to survive pesticide exposures, and combining compatible pesticides and natural enemies – are yet to be employed as fully as possible, or even as fully as practical. Hoy (1989) suggested that a reordering of priorities

was needed if biological control is to be integrated into agricultural IPM systems, and that chemical control should not be viewed as the central component around which other tactics must fit. Selective modes of action of some new pesticides provide opportunities that were not available in the past, and transgenic plants produce toxins that are selective as well. Nonetheless, a balanced, integrated approach to arthropod pest management seems little closer to reality than it was nearly two decades ago when Hoy (1989) suggested the need for new priorities. Transgenic crops seem to have become the new application method for insecticides, but integrating their use with natural enemies seems far beyond the vision of the pest control mindset that again seems to govern current agriculture. Challenges still loom large for true integration in modern-day IPM.

References Al-Deeb, M. A. & Wilde, G. E. (2003). Effect of Bt corn expressing the Cry3Bb1 toxin for corn rootworm control on aboveground nontarget arthropods. Environmental Entomology, 32, 1164–1170. Al-Deeb, M. A., Wilde, G. E., Blair, J. M. & Todd, T. C. (2003). Effect of Bt corn for corn rootworm control on nontarget soil microarthropods and nematodes. Environmental Entomology, 32, 859–865. Andow, D. A. & Hilbeck, A. (2004). Science-based risk assessment for nontarget effects of transgenic crops. BioScience, 54, 637–649. Barbara, K. A. & Buss, E. A. (2005). Integration of insect parasitic nematodes (Rhabditida: Steinernematidae) with insecticides for control of pest mole crickets (Orthoptera: Gryllotalpidae: Scapteriscus spp.). Journal of Economic Entomology, 98, 689–693. Bernard, M. B., Horne, P. A. & Hoffman, A. A. (2004). Developing an ecotoxicological testing standard for predatory mites in Australia: acute and sublethal effects of fungicides on Euseius victoriensis and Galendromus occidentalis (Acarina: Phytoseiidae). Journal of Economic Entomology, 97, 891–899. Boyd, M. L. & Boethel, D. J. (1998). Susceptibility of predaceous hemipteran species to selected insecticides used on soybeans in Louisiana. Journal of Economic Entomology, 91, 401–409. Brinkman, M. A. & Gardner, W. A. (2001). Use of diatomaceous earth and entomopathogen combinations against the red imported fire ant (Hymenoptera: Formicidae). Florida Entomologist, 84, 740–741.




Brown, G. C. (1997). Simple models of natural enemy action and economic thresholds. American Entomologist, 43, 117–124. Brust, G. E., Stinner, B. R. & McCartney, D. A. (1985). Tillage and soil insecticide effects on predator-black cutworm (Lepidoptera: Noctuidae) interactions in corn agroecosystems. Journal of Economic Entomology, 78, 1389–1392. Clark, J. M. & Yamaguchi, I. (2002). Scope and status of pesticide resistance. In Agrochemical Resistance: Extent, Mechanism, and Detection, eds. J. M. Clark & I. Yamaguchi, pp. 1–22. Washington, DC: American Chemical Society. Clausen, C. P. (1956). Biological Control of Insect Pests in the Continental United States, Technical Bulletin No. 1139. Washington, DC: US Department of Agriculture. Cloyd, R. A. & Dickinson, A. (2006). Effect of insecticides on mealybug destroyer (Coleoptera: Coccinellidae) and parasitoid Leptomastix dactylopii (Hymenoptera: Encyrtidae), natural enemies of citrus mealybug (Homoptera: Pseudococcidae). Journal of Economic Entomology, 99, 1596–1604. Conway, H. E., Steinkraus, D. C., Ruberson, J. R. & Kring, T. J. (2006). Experimental treatment threshold for cotton aphid (Homoptera: Aphididae) using natural enemies in cotton. Journal of Entomological Science, 41, 361– 373. Croft, B. A. (1990). Arthropod Biological Control Agents and Pesticides. New York: John Wiley. Croft, B. A. & Brown, A. W. A. (1975). Responses of arthropod natural enemies to insecticides. Annual Review of Entomology, 20, 285–334. Croft, B. A. & Meyer, R. H. (1973). Carbamate and organophosphorus resistance patterns in populations of Amblyseius fallacis. Environmental Entomology, 2, 691– 695. Curtis, J. E. & Horne, P. A. (1995). Effect of chlorpyrifos and cypermethrin applications on non-target invertebrates in a conservation-tillage crop. Australian Journal of Entomology, 34, 229–231. Davis, D. W. (1974). Parasite–prey ratios among alfalfa weevil larvae in northern Utah. Environmental Entomology, 3, 1031–1032. Desneux, N., Fauvergue, X., Dechaume-Moncharmont, F. et al. (2005). Diaterella rapae limits Myzus persicae populations after application of deltamethrin in oilseed rape. Journal of Economic Entomology, 98, 9–17. Egel, D., Lam, F., Foster, R. et al. (2007). Midwest Vegetable Production Guide for Commercial Growers, 2007. West Lafayette, IN: Purdue University.

Ericsson, J. D., Kabaluk, J. T., Goettel, M. S. & Myers, J. H. (2007). Spinosad interacts synergistically with the insect pathogen Metarhizium anisopliae against the exotic wireworms Agriotes lineatus and Agriotes obscurus (Coleoptera: Elateridae). Journal of Economic Entomology, 100, 31–38. Fernandez, O. A., Wright, R. J. & Mayo, Z. B. (1998). Parasitism of greenbugs (Homoptera: Aphididae) by Lysiphlebus testaceipes (Hymenoptera: Braconidae) in grain sorghum: implications for augmentative biological control. Journal of Economic Entomology, 91, 1315– 1319. Furlong, M. J. & Groden, E. (2001). Evaluation of synergistic interactions between Colorado potato beetle (Coleoptera: Chrysomelidae) pathogen Beauveria bassiana and the insecticides imidacloprid and cyromazine. Journal of Economic Entomology, 94, 344–356. Furlong, M. J., Zu-Hua, S., Yin-Quan, L. et al. (2004). Experimental analysis of the influence of pest management practices on the efficacy of an endemic arthropod natural enemy complex of the diamondback moth. Journal of Economic Entomology, 97, 1814–1827. Giles, K. L., Jones, D. B., Royer, T. A., Elliott, N. C. & Kindler, S. D. (2003). Development of a sampling plan in winter wheat that estimates cereal aphid parasitism levels and predicts population suppression. Journal of Economic Entomology, 96, 975–982. Gonzalez, D. & Wilson, L. T. (1982). A food-web approach to economic thresholds: a sequence of pests/predaceous arthropods on California cotton. Entomophaga, 27, 31–43. Grafton-Cardwell, E. E. & Hoy, M. A. (1986). Genetic improvement of common green lacewing, Chrysoperla carnea (Stephens) (Neuroptera: Chrysopidae): selection for carbaryl resistance. Environmental Entomology, 15, 1130–1136. Hagen, K. S. & McMurtry, J. A. (1979). Natural enemies and predator–prey ratios. In Biological Control and Insect Management, Bulletin No. 1191, eds. D. W. Davis, S. C. Hoyt, J. A. McMurtry & M. T. Aliniazee, pp. 28–40. Riverside, CA: University of California, Agricultural Experiment Station. Hagley, E. A. C. & Laing, J. E. (1989). Effect of pesticides on parasitism of artificially distributed eggs of the codling moth, Cydia pomonella (Lepidoptera: Tortricidae), by Trichogramma spp. (Hymenoptera: Trichogrammatidae). Proceedings of the Entomological Society of Ontario, 120, 25–33. Hassan, S. A. (1969). Observations on the effect of insecticides on coleopterous predators of Erioischia brassicae (Diptera: Anthomyiidae). Entomologia Experimentalis et Applicata, 12, 157–168.


Hassan, S. A. & Oomen, P. A. (1985). Testing the side effects of pesticides on beneficial organisms by OILP working party. In Biological Pest Control: The Glasshouse Experience, eds. N. W. Hussey & N. Scopes, pp. 145–152. Ithaca, NY: Cornell University Press. Higley, L. G. & Wintersteen, W. K. (1992). A novel approach to environmental risk assessment of pesticides as a basis for incorporating environmental costs into economic injury levels. American Entomologist, 38, 34–39. Hoffman, M. P., Wilson, L. T., Zalom, F. G. & Hilton, R. J. (1990). Parasitism of Heliothis zea (Lepidoptera: Noctuidae) eggs: effect on pest management decision rules for processing tomatoes in the Sacramento Valley of California. Environmental Entomology, 19, 753–763. Hoffman, M. P., Wilson, L. T., Zalom, F. G. & Hilton, R. J. (1991). Dynamic sequential sampling plan for Helicoverpa zea (Lepidoptera: Noctuidae) eggs in processing tomatoes: parasitism and temporal patterns. Environmental Entomology, 20, 1005–1012. Hollingsworth, R. G., Steinkraus, D. C. & McNew, R. W. (1995). Sampling to predict fungal epizootics in cotton aphids (Homoptera: Aphididae). Environmental Entomology, 24, 1414–1421. Horn, D. J. (1971). The relationship between a parasite, Tetrastichus incertus (Hymenoptera: Eulophidae), and its host the alfalfa weevil, Hypera postica (Coleoptera: Curculionidae), in New York. Canadian Entomologist, 103, 83–94. Hoy, M. A. (1985). Recent advances in genetics and genetic improvement of the Phytoseiidae. Annual Review of Entomology, 30, 345–370. Hoy, M. A. (1989). Integrating biological control into agricultural IPM systems: reordering priorities. In Proceedings, National Integrated Pest Management Symposium/Workshop, ed. E. H. Glass, pp. 41–57. Geneva, NY: New York State Agricultural Experiment Station, Cornell University. Hoy, M. A. (1995). Multitactic resistance management: an approach that is long overdue? Florida Entomologist, 78, 443–451. Hoy, M. A. (2000). Transgenic arthropods for pest management programs: risks and realities. Experimental and Applied Acarology, 24, 463–495. Hoy, M. A. (2006). Evaluating potential risks of transgenic arthropods for pest management programmes. In Status and Risk Assessment of the Use of Transgenic Arthropods in Plant Protection, Proceedings of the FAO/IAEA Programme of Nuclear Techniques in Food and Agriculture, International Atomic Energy Association TECDOC-1483, pp. 121–144. Rome, Italy: Food and Agriculture Organization of the United Nations.

Hoy, M. A. & Cave, F. E. (1988). Guthion-resistant strain of the walnut aphid parasite. California Agriculture, 42(4), 4–6. Hull, L. A. & Beers, E. H. (1985). Ecological selectivity: modifying chemical control practices to preserve natural enemies. In Biological Control in Agricultural IPM Systems, eds. M. A. Hoy & D. C. Herzog, pp. 103–122. Orlando, FL: Academic Press. Insecticide Resistance Action Committee (2006). IRAC Mode of Action Classification Version 5.2. Available at Classification_v5.2.pdf. James, R. R. & Elzen, G. W. (2001). Antagonism between Beauveria bassiana and imidacloprid when combined for Bemisia argentifolii (Homoptera: Aleyrodidae) control. Journal of Economic Entomology, 94, 357–361. Jones, C. J. & Weinzierl, R. A. (1997). Geographical and temporal variation in pteromalid (Hymenoptera: Pteromalidae) parasitism of stable fly and house fly (Diptera: Muscidae) pupae collected from cattle feedlots in Illinois. Environmental Entomology, 26, 421–432. Kovach, J., Petzold, C., Degni, J. & Tette, J. (1992). A Method to Measure the Environmental Impact of Pesticides, New York Food and Life Sciences Bulletin No. 139. Geneva, NY: Cornell University New York State Agricultural Experiment Station Available at http:// Landis, D. A., Wrattan, S. D. & Gurr, G. M. (2000). Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology, 45, 175–201. L¨ ovei, G. L. & Arpaia, S. (2005). The impact of transgenic plants on natural enemies: a critical review of laboratory studies. Entomologica Experimentalis et Applicata, 114, 1–14. Li, J. & Hoy, M. A. (1996). Adaptability and efficacy of transgenic and wild-type Metaseiulus occidentalis (Acari: Phytoseiidae) compared as part of a risk assessment. Experimental and Applied Acarology, 20, 563–573. Maagd, R. A., Bravo, A. & Crickmore, N. (2001). How Bacillus thuringiensis has evolved specific toxins to colonize the insect world. Trends in Genetics, 17, 193–199. Medina, P., Smagghe, G., Budia, F., Tirry, L. & Vi˜ nuela, E. (2003). Toxicity and absorption of azadirachtin, diflubenzuron, pyriproxyfen, and tebufenozide after topical application in predatory larvae of Chrysoperla carnea (Neuroptera: Chrysopidae). Environmental Entomology, 32, 196–203. Musser, F. E., Nyrop, J. P. & Shelton, A. M. (2006). Integrating biological and chemical control in decision making: European corn borer (Lepidoptera: Crambidae)




control in sweet corn as an example. Journal of Economic Entomology, 99, 1538–1549. Naranjo, S. E. & Hagler, J. R. (1998). Characterizing and estimating the impact of heteropteran predation. In Predatory Heteroptera: Their Ecology and Use in Biological Control, Thomas Say Symposium Proceedings, eds. M. Coli & J. Ruberson, pp. 170–197. Lanham, MD: Entomological Society of America. Naranjo, S. E., Head, G. P. & Dively, G. P. (2005). Field studies assessing arthropod non-target effects in Bt transgenic crops: Introduction. Environmental Entomology, 34, 1178–1180. National Research Council (1993). Pesticides in the Diets of Infants and Children. Washington, DC: National Academy of Sciences, National Research Council, Board on Agriculture and Board on Environmental Studies and Toxicology, Committee on Pesticides in the Diets of Infants and Children. National Research Council (2000). The Future Role of Pesticides in US Agriculture. Washington, DC: National Academy of Sciences, National Research Council, Board on Agriculture and Natural Resources and Board on Environmental Studies and Toxicology, Commission on Life Sciences, Committee on the Future Role of Pesticides in US Agriculture. Newsom, L. D., Smith, R. F. & Whitcomb, W. H. (1976). Selective pesticides and selective use of pesticides. In Theory and Practice of Biological Control, eds. C. Huffaker & P. Messenger, pp. 565–591. New York: Academic Press. Niemczyk, E., Olszak, R., Miszczayk, M. & Bakowski, G. (1974). Effectiveness of Some Predaceous Insects in the Control of Phytophagous Mites and Aphids on Apple Trees, Annual Report. Skierniewice, Poland: Research Institute of Pomology. Nyrop, J. P. & van der Werf, W. (1994). Sampling to predict and monitor biological control. In Handbook of Sampling Methods for Arthropod Pests in Agriculture, eds. L. Pedigo & D. G. Buntin. Boca Raton, FL: CRC Press. Ostlie, K. R., Hutchinson, W. D. & Hellmich, R. L. (eds.) (1997). Bt Corn and European Corn Borer, NCR Publication No. 602. St Paul, MN: University of Minnesota. Pielou, D. P. & Glasser, R. F. (1952). Selection for DDT resistance in a beneficial insect parasite. Science, 115, 117–118. Quintela, E. D. & McCoy, C. W. (1998). Synergistic effect of imidacloprid and two entomopathogenic fungi on the behavior and survival of larvae of Diaprepes abbreviatus (Coleoptera: Curculionidae) in soil. Journal of Economic Entomology, 91, 110–122. Ragsdale, D. W., McCornack, B. P., Venette, R. C. et al. (2007). Economic threshold for soybean aphid

(Hemiptera: Aphididae). Journal of Economic Entomology, 100, 1258–1267. Ridgway, R. L. & Jones, S. L. (1969). Inundative release of Chrysopa carnea for control of Heliothis in cotton. Journal of Economic Entomology, 62, 177–180. Ripper, W. E., Greenslade, R. M. & Hartley, G. S. (1951). Selective insecticides and biological control. Journal of Economic Entomology, 44, 448–459. Rosenheim, J. A. & Hoy, M. A. (1988). Genetic improvement of a parasitoid biological control agent: artificial selection for insecticide resistance in Aphytis melinus (Hymenoptera: Aphelinidae). Journal of Economic Entomology, 81, 1161–1173. Rothwangi, K. B., Cloyd, R. A. & Wiedenmann, R. N. (2004). Effects of growth regulators on citrus mealybug parasitoid Leptomastix dactylopii (Hymenoptera: Encyrtidae). Journal of Economic Entomology, 97, 1239– 1244. Rytter, J. L. & Travis, J. W. (2006). Pennsylvania Tree Fruit Production Guide, 2006–07 edn. University Park, PA: Pennsylvania State University. Available at Shelton, A. M., Zhao, J. Z. & Roush, R. T. (2002). Economic, ecological, food safety, and social consequences of the deployment of Bt transgenic plants. Annual Review of Entomology, 47, 845–881. Smith, R. F. & Hagen, K. S. (1959). Impact of commercial insecticide treatments. Hilgardia, 29, 131–154. Stark, J. D. & Banks, J. E. (2003). Population-level effects of pesticides and other toxicants on arthropods. Annual Review of Entomology, 48, 505–519. Stark, J. D., Vargas, R. & Miller, N. (2004). Toxicity of spinosad in protein bait to three economically important tephritid fruit fly species (Diptera: Tephritidae) and their parasitoids (Hymenoptera: Braconidae). Journal of Economic Entomology, 97, 911–915. Steffey, K. L. & Gray, M. E. (2007). Insect pest management for field and forage crops. In 2007 Illinois Agricultural Pest Management Handbook, ed. S. Bissonnette, pp. 1–19. Urbana, IL: University of Illinois Extension. Stern, V. M. & van den Bosch, R. (1959). Field experiments on the effects of insecticides. Hilgardia, 29, 103–130. Stern, V. M., Smith, R. F., Van Den Bosch, R. & Hagen, K. S. (1959). The integrated control concept. Hilgardia, 29, 81–101. Tamaki, G. & Long, G. E. (1978). Predator complex of the green peach aphid on sugar beets: expansion of the predator power and efficacy model. Environmental Entomology, 7, 835–842. Tamaki, G., McGuire, J. V. & Turner, J. E. (1974). Predator power and efficacy: a model to evaluate their impact. Environmental Entomology, 3, 625–630.


Teetes, G. L. (1994). Adjusting crop management recommendations for insect-resistant crop varieties. Journal of Agricultural Entomology, 11, 191–200. Theiling, K. M. & Croft, B. A. (1988). Pesticide sideeffects on arthropod natural enemies: a database summary. Agriculture, Ecosystems and Environment, 21, 191–218. Tillman, P. G. & Mullinix, B. G. Jr. (2004). Comparison of susceptibility of pest Euschistus servus and predator Podisus maculiventris (Heteroptera: Pentatomidae) to selected insecticides. Journal of Economic Entomology, 97, 800–806. Turpin, F. T. & A. C. York. (1981). Insect management and the pesticide syndrome. Environmental Entomology, 10, 567–572. Weinzierl, R. & Cloyd, C. (2007). Insect pest management for commercial vegetable crops. In 2007 Illinois Agricultural Pest Management Handbook, ed. S. Bissonnette, pp. 199–235. Urbana, IL: University. of Illinois Extension.

Weinzierl, R. & Henn, T. (1991). Alternatives in Insect Management: Biological and Biorational Approaches, North Central Regional Publication No. 401. Urbana, IL: University of Illinois Extension. Available at∼vista/abstracts/aaltinsec.html. Weires, R. W. (1980). Predatory Phytoseiid Mite, Amblyseius fallacis (Garman), Tree Fruit IPM Insect Identification Sheet No. 7. Geneva, NY: Cornell University New York State Agricultural Experiment Station. Whitten, M. J. & Hoy, M. A. (1999). Genetic improvement and other genetic considerations for improving the efficacy and success rate of biological control. In Principles and Applications of Biological Control, eds. T. S. Bellows Jr. & T. W. Fisher, pp. 271–296. San Diego, CA: Academic Press. Wilkinson, T. K., Landis, D. A. & Gut, L. J. (2004). Parasitism of obliquebanded leafroller (Lepidoptera: Tortricidae) in commercially managed Michigan apple orchards. Journal of Economic Entomology, 97, 1524– 1530.


Chapter 15

Pesticide resistance management Casey W. Hoy Good pesticide resistance management is just good IPM. Managing pesticide resistance means using the chemistry with enough restraint, and enough understanding of its role in an agroecosystem, to sustain that use. One interpretation is to maximize the number of applications that result in close to 100% control. To achieve this goal means optimizing the trade-off between the selection pressure exerted by use of the chemistry and the control benefits in terms of rapid reduction of pest population density. Too much population reduction too quickly leads quickly to resistance, whereas too much restriction of use denies economically justified control. The hope is that the chosen pattern of use provides many more effective applications than any alternative pattern, ideally an infinite number of effective applications. Models of resistance development, however, typically predict a finite number of effective applications, followed by a rapid increase in resistance gene frequency and rather sudden control failure. Experience in agroecosystems has generally been consistent with model predictions, for example resistance within a few generations to a long list of insecticides in Colorado potato beetle (Leptinotarsa decemlineata) on Long Island (Forgash, 1985; Mota-Sanchez et al., 2006). Agroecosystems include both people and the land, and their functioning needs to be consid-

ered in light of this combination and at large spatial extents and long time-frames. Pesticide resistance management programs that have taken a comprehensive and cooperative approach to maximizing the number of effective applications have been in place for about 20 years. A comprehensive strategic and theoretical analysis appeared in a National Research Council study (Glass, 1986) at a time when the first comprehensive and nationwide resistance management program was beginning in Australia (Forrester & Cahill, 1987), although pesticide resistance has been recognized and studied for a much longer period of time (e.g. Georghiou, 1972). So far, what has been sustained is an industrial cycle of innovation, therapeutic product use and product obsolescence because of resistance, all based on users who are heavily dependent on pesticides as economically valuable, therapeutic solutions to pest problems. One might ask, however, if sustaining reliance on a singular albeit profitable therapeutic solution to high pest population density is the goal of IPM (see Chapter 1). An alternative interpretation of pesticide resistance management would be to prevent it completely. Meeting this objective could be considerably more difficult than maximizing the number of effective applications. It would require finding trade-offs that lead to evolutionary dead ends

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


for resistant insects, as sometimes documented in natural systems, rather than the ultimate success that resistant insects typically have achieved in pesticide treated environments. Alternatively, it could push insect adaptation in other directions, such as adaptation to non-crop hosts. The impact of pesticide chemistry and use pattern on the full range of life history traits, trophic relationships and other control tactics would need to be explored and understood in detail. We currently tend to understand little of this biology and ecology compared with our knowledge of a use pattern that kills pests. A pesticide use pattern that truly could be sustained indefinitely may not be geared to providing quick and complete control of pests, but rather may be geared towards shifting their population dynamics and host plant relationships in much more subtle ways, adding to existing ecosystem services rather than replacing them. Such shifts are theoretically the role of a single tactic in IPM, to be combined with many other similar tactics to prevent economic injury. But in practice, this level of integration of chemical controls is rarely achieved and is not considered to be pesticide resistance management. The economic benefits of such a use pattern may be too diffuse to support the level of industrial activity needed to produce and distribute the pesticide in the first place. Nevertheless, use patterns that avoid resistance entirely would be more consistent with the theory of IPM, as discussed in Chapter 1, and should be the ultimate goal of pesticide resistance management as well whether or not it is currently achievable with available chemistries, technologies for their use and the agroecosystems in which this use takes place. This chapter will outline first the biology and ecology of pesticide resistance and second the extent to which resistance management has improved the resilience of agroecosystems. Insecticide resistance in crop production will be the focus, although many of the principles could apply to other commodities and public health as well as weeds and pathogens. Although the institutional responses to resistance have become well organized and entrenched, opportunities for improving the resilience of agroecosystems may still be available.

15.1 Genotypes and phenotypes in individuals, populations and ecosystems 15.1.1 Evolutionary underpinnings of pesticide resistance Herbivores have been dealing with chemical defenses in host plants, a basis of host plant resistance, for a long time. Although plant defenses are ubiquitous, a number of trade-offs act to keep toxin concentrations very heterogeneous in natural systems at spatial scales from within leaves to across continents (Berenbaum, 1995; Hoy et al., 1998). Furthermore, variation in environmental conditions can lead to variation in chemical defenses in plants (Herms & Mattson, 1992). Physiological adaptation to plant toxins does occur in natural systems, particularly in specialist herbivores for which the ultimate physiological adaptation is complete immunity and storage or sequestering of the toxic compounds from the host plant to provide chemical defense to the herbivore itself. In these cases, however, the chemical defense can still play a role in protection from other herbivores and microbes. Expression of naturally occurring plant chemical defenses is shaped by multiple herbivore guilds, leading to complex and varying chemical profiles and responses (Adler & Kittelson, 2004). In most cases, however, stability in plant defense is afforded by the heterogeneous distributions resulting from trade-offs, including tradeoffs between resistance to multiple threats such as pathogens and herbivores (Felton & Korth, 2000). Theories behind these trade-offs generally require a cost to the plant for production of the chemical defense. Initial attempts to quantify these costs suggested that they are small or non-existent (Bergelson & Purrington, 1996). More recent metaanalysis, however, indicates that resistance costs to plants are much more common than previously estimated and accrue from both the direct cost of resource allocation within plants and ecological interactions between plants and their environment (Koricheva, 2002; Strauss et al., 2002). Two kinds of responses have been documented in insect herbivores to heterogeneous distributions of naturally occurring toxins in plants.




Physiological adaptation has been documented, but often constrained to an apparently stable pattern of defense and adaptation between the plants and insects (e.g. Berenbaum et al., 1986; Berenbaum & Zangerl, 1998; Nitao, 1995). Far more common, however, are behavioral adaptations to plant toxins that shift feeding away from higher concentration and toward lower concentration (Hoy et al., 1998). The concentrations and spatial scales of pesticide treated and untreated areas typically are not intended to allow behavioral responses to result in survival of the pests, despite the fact that avoidance of treated areas can often accomplish the objectives of IPM. The typical adaptive response of pest populations in managed systems is adaptation by physiological resistance (Mota-Sanchez et al., 2002). Despite long-standing recognition (Berenbaum, 1995) of these differences between deployment of defensive chemistry in agriculture and in nature, and associated problems including resistance to pesticides, we persist in using pesticides very differently from the way chemical defenses have evolved in plants to a relatively stable arrangement.

15.1.2 Mechanisms and genes for pesticide resistance Although many investigations of insecticide resistance have focused on finding the gene (singular) responsible, a wide range of genetic mechanisms responsible for reduced penetration or absorption, detoxification, sequestering or target site insensitivity have been identified for many if not most insecticide chemistries, and these have been reviewed previously (Georghiou, 1972; Denholm & Rowland, 1992; Brown, 1996). These mechanisms are not mutually exclusive and frequently are found in combination, often with uncharacterized modifiers or genes of “minor effect.” The modifier genes may contribute only small amounts of variation in tolerance, but their potential correlation with other traits or contribution to continued adaptation, reduction of fitness effects, etc. should not be overlooked. Fitness effects in particular are typically presented as if static, but may remain under selection in a resistance management program that relies on them in repeated cycles of selection. Fitness effects also may not be extensive enough to provide sufficient suppres-

sion of resistance gene frequencies in a practical pest management program. For example, rotating between pyrethroids and other chemistries in one of the first comprehensive resistance management programs in Australian cotton resulted in a “sawtooth pattern” of resistance frequency but with a long-term incline toward greater and greater frequencies of pyrethroid resistance (Forrester & Bird, 1996).

15.1.3 Phenology, dispersal and other life history characteristics and the spatial and temporal scales of resistance When resistance occurs it typically arises at a particular spatial and temporal scale, which depends on the life history characteristics of the pest, particularly the spatial extent of gene flow, and the pesticide use pattern. Cases of resistance can arise over areas from individual fields or farms to entire regions. Following initial occurrence it will typically expand geographically, but the rate of expansion will be dependent upon dispersal patterns of the pest at multiple scales. The temporal scales of resistance evolution are determined by such factors as the selection pressure, gene flow, stability of the resistance, presence of cross-resistance patterns and associated pesticide use, number of mechanisms and other correlated traits (Georghiou, 1986).

15.2 Navigating the everywhere/always – nowhere/never continuum In general, resistance management means a reduction in use of the pesticide from what could theoretically be beneficial in the short term, i.e. if resistance was not of concern. A restriction in use can come in one or more of three ways: the concentration used, the area treated or the frequency of application. A simple set of rules to restrict pesticide use can be devised but the result of using them will be very complex due to variable spatial arrangements, time-frames and pest population dynamics. In practice, however, these are connected in ways that need to be considered


Fig. 15.1 Spatial and temporal scales of common resistance management tactics. Producers dominate the decisions and implementation within the shaded area whereas institutions dominate the management of tactics outside of the shaded area.

carefully and matched with the spatial and temporal scales at which resistance could develop (Fig. 15.1). For example, temporal alternations or rotations operated independently within small contiguous areas like individual farms amount to spatial mosaics over larger areas like farm landscapes. Refuge areas are likely to be moved from season to season, to avoid continued buildup of pest populations in a particular area. The sequence of treatments within a particular refuge area, therefore, scales up to an alternation or rotation over longer time-frames. The larger the area treated, the less likely will the timing be optimal for all pests species within the area, given variation in microclimate, oviposition and development. But the smaller the area treated, the less likely exposure will be confined to just the pests in the treated area as movement across the treated area boundaries is more likely to occur. Weathering of residues and dilution of concentrations by plant growth begin immediately after an application, so the concentration of pesticide is constantly in flux. Despite the need to keep resistance management programs simple, the situation in the field is likely to remain complex.

15.2.1 Prevention or cure? Managing resistance can take the form of preventing it from occurring in the first place or decreasing the frequency of already resistant pests in a population. Theoretical studies suggest that prevention will only be possible if the preventive strategy is in place when resistance gene frequen-

cies are very low (Tabashnik & Croft, 1982), but at this point the genetics of resistance are usually unknown or at best a guess. Prevention is generally preferred, but reliance on pesticides for economically beneficial, rapid and thorough population reduction results in use patterns that make resistance very likely, along with a hope that it will be unstable and manageable. Selectively reducing the frequency of resistant pests or increasing the frequency of susceptible pests requires some behavioral or biochemical means of selection. Human pathogen resistance to antibiotics has provided both theoretical and empirical results that have been borrowed in developing approaches for managing pesticide resistance. In particular the question of low versus high dose, or the attention given to eradication of the pathogen with each use, has been applied to resistance management for agricultural pests. The focus on leaving no survivors leads to choices for how aggressively to attack the pest population, using single or multiple pesticides simultaneously. The more aggressive high dose and pesticide mixture approaches both maximize selection pressure and minimize initial numbers of resistant organisms, potentially increasing the effectiveness of a spatial refuge (Caprio, 1998). The specific tactics available to manage resistance in pest populations have been reviewed comprehensively (Roush, 1989; Denholm & Rowland, 1992). In addition to variations on restricted use of a given pesticide in space or time, they include the use of pesticides, or mixtures of multiple




pesticides and/or synergists, that are known to be effective for a particular resistant population of pests. The stated goal in these cases, however, has been to delay rather than prevent resistance. Testing these various strategies is difficult because they take place at the landscape scale and over relatively long periods of time. One of the few large-scale field studies undertaken took advantage of isolated pockets along the coast of Mexico to examine mosaics, rotations and single insecticides (i.e. the first stage of an alternation. There, all of these strategies led to elevated levels of resistance in mosquitos within one year (Hemingway et al., 1997). The opportunities for prevention of resistance deserve some additional attention, despite the obvious change in the entire approach to pesticide use that would be required.

15.2.2 Preserving susceptible pests Options for improving the survival of susceptible pests in a treated environment include reducing the concentration of pesticides, and moving from always and everywhere toward never and nowhere on the use continuum. In practice, strategies have focused on some variation on either windows in time or refuges in space, or both, where pesticides are not used at all. An additional and rarely considered possibility is the use of concentrations that are too low to kill even susceptible insects. Such a strategy relies on effects other than direct mortality, such as behavioral responses or altered development times, to protect against pest damage. Spatial heterogeneity in pesticide concentrations typically has taken the form of high doses in treated areas, meant to kill insects that are heterozygous for resistance, and completely untreated refuge areas to support populations of susceptible insects (see Chapter 19 for specific examples with transgenic crops). The untreated areas can occur naturally, particularly for polyphagous pests with suitable abundance of alternate hosts. In the absence of a naturally occurring refuge, a design for a refuge arrangement can be challenging. The hope is that enough susceptible insects will disperse into treated areas to ensure production of heterozygotes from any resistant pests present, which will then be killed by the insecticide. The corollary is a hope that resistant pests do not move into the refuge and

interbreed there, creating more homozygotes. Knowing or estimating gene flow relative to scale of treated and untreated areas can permit estimation of a size for treated areas that prevents the accumulation of resistant genotypes, based on a number of assumptions including a single gene for resistance and stable fitness effects (Lenormand & Raymond, 1998). In fact, the assumptions needed for an effective refuge are fairly restrictive (Carri`ere & Tabashnik, 2001). Lots of things can go wrong including insufficient gene flow or non-random mating (Caprio, 2001). The common assumption regarding low pesticide concentrations is that they could permit survival of heterozygotes and make resistance effectively dominant (see Chapter 19). Certainly exposure to a low and uniform concentration of a pesticide, a concentration that selectively kills susceptible insects and permits survival of resistant ones, will have a very predictable result. Low concentrations, however, do not have to be spatially or temporally uniform, and adaptation in an insect need not be entirely physiological. Insects can respond either behaviorally or physiologically to low and spatially heterogeneous concentrations of pesticides (Hoy et al., 1998), and behavioral responses tend to take place at concentrations well below those required to kill even susceptible insects. When a pesticide is applied to a surface, particularly a target like a crop canopy with rather complex architecture, the state of current application technology guarantees that it will be heterogeneous at a scale to which insects can respond behaviorally. Breakdown by UV light and weathering and dilution by growth occur immediately, leading to a very ephemeral and patchy deposit on plant surfaces, and even within plant tissues for systemic pesticides. Understanding how insects respond and adapt behaviorally as well as physiologically can be very important. Low and heterogeneous concentrations of pesticides can select for susceptibility rather than resistance. Georghiou (1972) proposed that avoidance and tolerance of a pesticide should be negatively correlated in insect populations, i.e. the most behaviorally responsive, hypersensitive insects should also be the most susceptible. A similar trade-off is hypothesized for mammalian herbivores between tolerance and avoidance (Iason


& Villalba, 2006). Using simple models with a single gene for resistance and a single gene for avoidance, simulation results indicate that avoidance can result in very stable behavioral adaptation rather than physiological resistance (Gould, 1984). Particularly when behavioral responses are involved, however, single-gene assumptions are not likely to be valid. Using a variety of experimental techniques and examining a wide variety of arthropod species, a range of correlations between behavioral responsiveness and physiological tolerance have been recorded from negative to neutral to positive (Hoy et al., 1998). Empirical results describing the outcome of selection in a heterogeneously treated environment are very rare but intriguing results have been observed. For example, susceptible aphids survived better than moderately resistant aphids on treated plants presumably by moving to untreated parts of the plant canopy (Ffrench-Constant et al., 1988). Our work on diamondback moth (Plutella xylostella) using quantitative genetics methodology has demonstrated significant genetic variation in both tolerance and behavioral responsiveness to permethrin, and negative genetic correlations between these traits (Head et al., 1995a, b, c; Jallow & Hoy, 2005, 2006). We recently conducted a selection experiment (Jallow & Hoy, 2007) in greenhouse cages in which all life stages were exposed to one of three treatment regimes: untreated control, relatively uniform high concentration of permethrin (applied to the entire plant to runoff) and a relatively heterogeneous low concentration (applied only to the center leaves). Not surprisingly, the uniform high concentration resulted in a substantial increase in LC50 within 20 generations. However, the heterogeneous low dose resulted in a significant decrease in LC50 after 20 generations. Although it may take longer because of generally lower heritability for behavioral traits and indirect selection on behavior, it is possible to increase susceptibility to an insecticide in a treated environment. Furthermore, the behavioral response should be self-reinforcing in this case, as more susceptible diamondback moths were also more behaviorally responsive. The behavioral response resulted in avoidance of the treated parts of the plant, which were also the harvestable parts. The genetic cor-

relation can also be positive when toxin concentrations are high (Hoy & Head, 1995), in which case behavior could select indirectly for physiological resistance. Pest genetics and toxin distribution must be carefully matched to push selection in a direction that is advantageous to agriculture. Restricting pesticide use over time results in susceptible pests surviving either within or among generations. Control is typically most beneficial at critical times in the development of a crop, providing a focus or window in time when the pesticide can be used with the greatest benefit and least selection pressure. Logistical constraints to implementing windows and refuges can be very challenging (Forrester, 1990). Not the least among these challenges is that the restricted use of the pesticide often provides less immediate benefit to the producer than if used everywhere, constantly and at high concentration, i.e. attempted eradication rather than IPM.

15.2.3 Destroying resistant pests If resistance is to be cured, then some means of weeding out the resistant individuals in the population is needed. Options include very high doses, mixtures that overwhelm pests that are resistant to only part of the mixture, synergists that inhibit or overcome a particular resistance mechanism, mixtures with a negative cross-resistance pattern or some other means of selection based on the fitness effects of resistance (Roush, 1989; Denholm & Rowland, 1992). There is no guarantee that such mixtures or synergists exist for a given case of resistance. Barring insecticidal means of suppressing already resistant pests, the remaining means are the same suite of mortality factors that were available without the pesticide, and perhaps fitness effects of resistance. If these other mortality factors were sufficient for control in a given production system then pesticides would not be entirely necessary and probably would not have been used enough to result in resistance in the first place. Insect natural enemies and prey other than pest species tend to fare even more poorly than pests in treated environments. Pathogens and other mortality factors often depend on high pest population density. Thus, the reliance on therapeutic control tends to be self-reinforcing.




15.3 Agroecosystem-level responses to pesticide resistance Taken over the longer term and considering the social system accompanying pesticide use, pesticide resistance management fits a typical pattern of interplay between natural and social systems seen elsewhere in natural resource management, referred to as the resource management pathology (Holling et al., 2002). An immediate response to a serious challenge results in rapid and visible results, but sets in place a social system that precludes alternatives and further adaptation. Essentially the response and associated infrastructure that arises in social institutions (government organizations such as the US Environmental Protection Agency, industry organizations such as the Insecticide Resistance Action Committee (IRAC), universities and consortia such as the European Network for the Management of Arthropod Resistance to Insecticides and Acaricides) lock a system in place (the pesticide product cycle) that constrains the options for a lasting solution. A common assumption for environmental management that is driven by industrial interests underlies this syndrome, that nature can be controlled with sufficient engineering and management control, including new product development in this case. This is despite the recognition over 20 years ago by the National Research Council’s Committee on Strategies for the Management of Pesticide Resistant Pest Populations (Glass, 1986) that “Control of pest populations by combining in cycles the use of old and novel chemical pesticides, as they become available, is unlikely to be a viable longterm strategy.” “Victory” in this form of “combat” against pesticide resistance doesn’t appear to be any more feasible than winning the coevolutionary arms race for a plant species (Kareiva, 1999). The resistance management strategies recommended by IRAC include a healthy dose of IPM, suggesting as usual that good IPM will eliminate the concern about resistance. But in practice, new insecticides are still released with a marketing strategy based on providing very clear and saleable benefits of the product, i.e. use the product only

when a pest population has exceeded threshold (or prophylactically as in the case of systemic insecticides such as neonicotinoids applied at planting) and then achieve as close to 100% control as possible. Transgenic crops are following the same path, with the high dose (and impressive results) strategy taken as a given despite documented violation of key resistance management assumptions and severe practical difficulty with implementation especially in an IPM context (Bates et al., 2005). Pesticide resistance is perceived by industry to be a part of doing business, a motivator for better product stewardship, and therefore a contributor to sustainable agriculture (Urech et al., 1997). These perceived benefits, however, rest upon the assumption that pesticides must continue in their current role in agriculture. For example, implementation of the 1996 US Food Quality Protection Act has required a very careful and deliberate search for effective replacements for any pesticides removed from commercial use. The technical community takes an expected role in pesticide resistance: scientific inquiry, skepticism, debate and thorough analysis of the topic from many angles. This debate has played out in the scientific literature over virtually every aspect of resistance management. The debate itself leads to a certain amount of resistance to change in the approach taken by industry and government, which will be influenced by political pressure in the absence of complete agreement in the scientific community. These debates continue (for good reason, they’re part of good science) despite early recognition that one can already put into practice a simple and useful strategy based on existing knowledge that will gain most of the benefit of a more thoroughly researched approach (Roush, 1989), particularly if the strategy maintains the ability to change and adapt (Forrester, 1990). Are we moving toward greater resilience in the face of the adaptive capability of pests or trading resilience for a series of short-term solutions to this long-term problem? In general, experience with environmental management leads to a number of relevant questions that could guide an evaluation of where we are headed (Berkes et al., 2003). As cycles of pesticide product release and resistance have been repeated, is there a social and


Fig. 15.2 The cycle of adaptive renewal associated with pesticide resistance, modified from Holling (2001).

ecological memory accumulating that is improving the capacity of agroecosystems to adapt to rapid change (the “back-loop in the cycle of ecological and social change caused by resistance”) (see Fig. 15.2)? As expected, resistance still occurs, e.g. Colorado potato beetle has recently become resistant to imidacloprid on Long Island, NY, with cross-resistance to the other neonicotinoids and possibly to spinosad (Mota-Sanchez et al., 2006), despite extensive experience with resistance in this pest. The institutional response is that it was monitored closely and documented quickly and farmers can now move on to the next chemistry. The system we have in place around resistance is one in which the resistance problems occur locally and progress to regional and national or international levels, whereas the solutions (the next product in line) flow in the opposite direction (Fig. 15.3). Four properties have been suggested to lead to sustainability and resilience (Folke et al., 2003), and these will be considered in light of current approaches to resistance management: (1) Learning to live with change and uncertainty Resistance management systems currently in place do a good job of living with change and uncer-

tainty in terms of assuming that resistance will occur and taking a proactive stance, but do a poor job of living with change and uncertainty in their assumption that resistance will be manageable (identifiable single recessive or incompletely dominant genes, fitness costs or negative cross-resistance patterns, etc.) and that replacement chemistry will be available when needed. (2) Nurturing diversity for reorganization and renewal The cycle of resistance and new chemistry introduction encourages innovation on an industrial scale. Competition against the first product of a very effective new chemistry, however, can be very difficult and may suppress innovation even more than pesticide regulation. Industry, regulatory agencies and university or government research organizations have become very well organized to respond to changes in pesticide product cycles. The monitoring and educational programs needed to implement industry-wide resistance management programs are being developed more quickly and consistently, and have been very important in the response to increasing resistance problems (McCaffery & Nauen, 2006). Resistance management strategies are




Fig. 15.3 The panarchy of the adaptive renewal cycle currently associated with pesticide resistance. Although resistance occurs at a local scale, the problems scale upwards to the national and international scale, whereas solutions flow from innovation in multinational corporations down to local scales.

heavily reliant upon cooperation among individual producers and in many cases this cooperation has been successfully gained. At more local scales, however, agroecosystems that are reliant enough on pesticides to make resistance management an important issue tend to be homogeneous biologically, having natural capital suppressed and replaced by pesticides. As a result, when resistance occurs a shift to more biologically based pest management tends to be very difficult. When no replacement chemistry is available, the commodity may simply be lost, placing severe economic stress on producers who generally are not gaining flexibility. Strategies that require refuges, however, provide some opportunity to introduce increased bio-

logical diversity within agroecosystems. Management of biological diversity in refuges, rather than just production of susceptible insects, is an opportunity that deserves further research. (3) Combining different types of knowledge for learning Experience with pesticide resistance management has accumulated at the institutional level. Institutional learning appears to have prepared industry, research and educational and regulatory agencies to address similar issues with the next new chemistry. However, this institutional level of organization tends to suppress rather than enhance innovation and inventiveness on the part of individual producers or users of pesticides, and even inhibit new players in the pesticide industry. The free market forces at play encourage simply using or even overusing pesticides for the sake of consistency in production, leaving little learning among users other than to follow the instructions of extension and industry in use of the next product in the cycle when resistance occurs.


(4) Creating opportunity for self-organization Pesticide resistance management efforts have to some extent achieved a self-organizing community with an ability to quickly adapt at the institutional level, with coordination by government agencies such as the US Environmental Protection Agency, the industrybased IRAC, commodity organizations like the US Cotton Board, universities and associated grower outreach programs. However, little self-organization is encouraged at more local scales given the suppression of human and natural capital described above. Chemistries and uses that foster and improve natural capital, rather than replacing it, may be the only means of re-engaging the producer and creating opportunities for selforganization at local scales, relieving some of the dependence on solutions from larger scales and creating more resilient agroecosystems.

15.4 Conclusions Much of the current focus on resistance management stems from the need to respond to recalcitrant problems in a highly evolved and efficient system of agriculture that is very dependent upon external inputs like pesticides. The scientific and technical community is expected to respond with answers and solutions. We can either proactively prevent such problems by redesigning agroecosystems based on our current science and technology or accept the system as is and continue responding. Given that pesticides will continue to be an important management tactic for the foreseeable future, finding ways to create evolutionary dead ends for resistant pests deserves much more attention. Because most major companies are shifting from development of chemistries to development of traits, the advances in this approach could play out in crop and animal genetics (see Chapter 19) rather than in pesticide application. Unfortunately, genetic engineering currently is continuing along the same high-dose, high-kill, resultsthat-sell path as pesticides have been on, based at least as much on the business need to compete with pesticides as on the scientific arguments for the high dose-refuge strategy of resistance man-

agement. Although no cases of resistance to Bacillus thuringiensis (Bt) crops have emerged so far, the selection pressure they impose on pest populations is such that physiological resistance would be the likely pest response, eventually. A social and ecological system with a cycle of product innovation and obsolescence that takes years to decades is well established, at least in more developed countries, and should be able to continue on its own. The resistance management side of this system should require less support from public sources for research and regulation over time, and the private sector would prefer the reduction as well (Thompson & Head, 2001), at least in regulation. Although pesticide resistance has been around for almost a century and the science to understand it began at the first observations, we’ve really been responding as a scientific, technical and agricultural community in a coordinated and serious fashion to pesticide resistance for about two decades. What appeared to be an exponential increase in reported cases of resistance in the early 1980s has been an approximately linear increase since, with most of the new reported cases repeating previously reported cases in new locations (Mota-Sanchez et al., 2002). Farmers continue to rely on pesticides for economic survival, resistance continues to result and the agroecosystems that produce resistance continue to have fewer farmers and lower prospects for economic survival among those who remain. A new direction described in the 2006 strategic plan of the International Rice Research Institute focuses on the well-being of rice farmers, rather than strictly on rice production. This perceptive approach recognizes that farmers can think long term and use sustainable practices when they have sufficient relief from the economic pressures of their production system. Desperation measures taken by farmers who do not have that freedom, on the other hand, tend to be self-reinforcing, perpetuating a cycle of unsustainable approaches and declining resilience. This same “vicious cycle” is at play in more developed countries with more industrial approaches to agriculture. With dependence on pesticides that exert complete control comes a self-reinforcing cycle of resistance and dependence on new pesticides.




Shifting the emphasis of resistance management to self-organizing systems at local scales is a difficult challenge but at least two opportunities deserve more consideration: (1) Engage the creativity of producers Resistance monitoring should be in the hands of the user community. Some progress has been made with offering simple test kits to producers but more attention and more simple monitoring tools would be worthwhile. If agricultural producers could gain more flexibility in use patterns within limits that prevent environmental, health and resistance risk, we could benefit from their creativity and innovation, but current pest management systems do not encourage this kind of innovation. The extension service can help with sharing information more rapidly and widely on what works at the farm level. (2) Better imitate the diversity in toxins found in natural systems Heterogeneity in toxins and toxin concentrations at multiple spatial scales is a foundation of the relative stability seen in natural systems. Pesticide application remains imperfect enough to inherently provide diversity in concentrations. Beyond the considerable attention given to a high dose and refuge strategy for transgenic crops, however, the rational design of spatial heterogeneity in toxin distributions has received far less attention than the attempt to avoid it with uniform high concentrations. Behavioral adaptation in arthropod pests can be a means of stabilizing crop defense rather than a loss of control, if accompanied by shifts in the location of feeding damage that decrease economic losses or shifts in life history traits that enhance the activity of natural enemies. Perhaps the focus of discovery should also be intensified for new chemistries or families of toxicants with inherent negative cross-resistance patterns. New application technology may provide opportunities to create more diverse toxin mixture distributions with chemical applications. For example, a relatively new double nozzle (Chapple et al., 1996, 1997) expresses small pesticide solution droplets into a stream of larger

plain water droplets that carry the pesticide droplets into a plant canopy with improved efficiency of delivery. Droplets from multiple solutions could theoretically be injected into the same water carrier droplet stream to generate complex deposits with individual droplets of different chemistries. If this heterogeneity can incorporate negative cross-resistance patterns then it could further stabilize pest adaptive responses (Pittendrigh & Gaffney, 2001; Pittendrigh et al., 2004). The final conclusion on this subject remains unchanged over the past two decades. Pesticide resistance management will be enhanced more by the multiple tactics (especially non-chemical) and their integration described in the rest of this book than in continued refinement of resistance management strategies in agroecosystems that do not yet live up to the preventive potential of true IPM. Defensive chemistry could still be part of this potential but with a rather different role than the silver bullet it currently represents. The design capability needed to fit new chemistry into this potential role will take commitment, painstaking research and detailed ecological understanding. Meanwhile, pesticide resistance management programs that focus on maintaining the impressive benefits of complete control with pesticides will maintain both reliance and dependence on them.

References Adler, L. S. & Kittelson, P. M. (2004). Variation in Lupinus arboreus alkaloid profiles and relationships with multiple herbivores. Biochemical Systematics and Ecology, 32, 371–390. Bates, S. L., Zhao, J. Z., Roush, R. T. & Shelton, A. M. (2005). Insect resistance management in GM crops, past, present and future. Nature Biotechnology, 23, 57– 62. Berenbaum, M. R. (1995). The chemistry of defense: theory and practice. Proceedings of the National Academy of Sciences of the USA, 92, 2–8. Berenbaum, M. R. & Zangerl, A. R. (1998). Chemical phenotype matching between a plant and its insect herbivore. Proceedings of the National Academy of Sciences of the USA, 95, 13 743–13 748.


Berenbaum, M. R., Zangerl, A. R. & Nitao, J. K. (1986). Constraints on chemical coevolution, wild parsnips and the parsnip webworm. Evolution, 40, 1215–1228. Bergelson, J. & Purrington C. B. (1996). Surveying costs of resistance in plants. American Naturalist, 148, 536– 558. Berkes, F., Colding, J. & Folke, C. (eds.) (2003). Navigating Social–Ecological Systems. Cambridge, UK: Cambridge University Press. Brown, T. (1996). Molecular Genetics and Evolution of Pesticide Resistance. Washington, DC: American Chemical Society. Caprio, M. (1998). Evaluating resistance management strategies for multiple toxins in the presence of external refuges. Journal of Economic Entomology, 91, 1021– 1031. Caprio, M. (2001). Source–sink dynamics between transgenic and non-transgenic habitats and their role in the evolution of resistance. Journal of Economic Entomology, 94, 698–705. Carri`ere, Y. & Tabashnik, B. (2001). Reversing insect adaptation to transgenic insecticidal plants. Proceedings of the Royal Society of London B, 268, 1475–1480. Chapple, A. C., Downer, R. A., Wolf, T. M., Taylor, R. A. J. & Hall, F. R. (1996). The application of biological pesticides: limitations and a practical solution. Entomophaga, 41, 465–474. Chapple, A. C., Wolf, T. M., Downer, R. A., Taylor, R. A. J. & Hall, F. R. (1997). Use of nozzle-induced air-entrainment to reduce active ingredient requirements for pest control. Crop Protection, 16, 323–330. Denholm, I. & Rowland, B. W. (1992). Tactics for managing pesticide resistance in arthropods, theory and practice. Annual Review of Entomology, 37, 91–112. Felton, G. W. & Korth, K. L. (2000). Trade-offs between pathogen and herbivore resistance. Current Opinion in Plant Biology, 3, 309–314. Ffrench-Constant, R. H., Clark, S. J. & Devonsire, A. L. (1988). Effect of decline of insecticide residues on selection for insecticide resistance in Myzus persicae Sulzer (Hemiptera: Aphididae). Bulletin of Entomological Research, 78, 19–30. Folke, C., Colding, J. & Berkes, F. (2003). Synthesis: building resilience and adaptive capacity in socialecological systems. In Navigating Social–Ecological Systems, eds. F. Berkes, J. Colding & C. Folke, pp. 352–387. Cambridge, UK: Cambridge University Press. Forgash, A. J. (1985). Insecticide resistance in the Colorado potato beetle. In Proceedings of 17th International Congress of Entomology, Research Bulletin No. 704, eds. D. N. Ferro & R. H. Voss, pp. 33–52. Amherst, MA: Mas-

sachusetts Agricultural Experiment Station, University of Massachusetts. Forrester, N. W. (1990). Designing, implementing and servicing an insecticide resistance management strategy. Pesticide Science, 28, 167–179. Forrester, N. W. & Bird, L. J. (1996). The need for adaption to change in insecticide resistance management strategies: the Australian experience. In Molecular Genetics and Evolution of Pesticide Resistance, ed. T. Brown, pp. 160–168. Washington, DC: American Chemical Society. Forrester, N. W. & Cahill, M. (1987). Management of insecticide resistance in Heliothis armigera (H¨ ubner) in Australia. In Combating Resistance to Xenobiotics: Biological and Chemical Approaches, eds. M. Ford, B. P. S. Khambay, D. W. Holloman & R. M. Sawicki, pp. 127–137. Chichester, UK: Ellis Horwood. Georghiou, G. P. (1972). The evolution of resistance to pesticides. Annual Review of Ecology and Systematics, 3, 133–168. Georghiou, G. P. (1986). The magnitude of the resistance problem. In Pesticide Resistance: Strategies and Tactics for Management, pp. 14–43. Washington, DC: National Academy Press. Glass, E. H. (ed.) (1986). Pesticide Resistance: Strategies and Tactics for Management. Washington, DC: Committee on Strategies for the Management of Pesticide Resistant Pest Populations, National Academy Press. Gould, F. (1984). Role of behavior in the evolution of insect adaptation to insecticides and resistant host plants. Bulletin of the Entomological Society of America, 30, 34–41. Head, G., Hoy, C. W. & Hall, F. R. (1995a). The quantitative genetics of behavioral and physiological response to a pyrethroid in diamondback moth. Journal of Economic Entomology, 88, 447–453. Head, G., Hoy, C. W. & Hall, F. R. (1995b). Permethrin droplets influence larval Plutella xylostella (Lepidoptera: Plutellidae) movement. Pesticide Science, 45, 271–278. Head, G., Hoy, C. W. & Hall, F. R. (1995c). Effects of direct and indirect selection on behavioral response to permethrin in larval diamondback moths. Journal of Economic Entomology, 88, 461–469. Hemingway, J., Penilla R. P., Rodriguez, A. D. et al. (1997). Resistance management strategies in malaria vector mosquito control: a large-scale field trial in southern Mexico. Pesticide Science, 51, 375–382. Herms, D. A. & Mattson, W. J. (1992). The dilemma of plants, to grow or defend. Quarterly Review of Biology, 67, 283–335.




Holling, C. S. (2001). Understanding the complexity of economic, ecological, and social systems. Ecosystems, 4, 390–405. Holling, C. S., Gunderson, L. H. & Ludwig, D. (2002). In quest of a theory of adaptive change. In Panarchy: Understanding Transformations in Human and Natural Systems, eds. L. H. Gunderson & C. S. Holling, pp. 3–22. Washington, DC: Island Press. Hoy, C. W. (1999). Colorado potato beetle resistance management strategies for transgenic potatoes. American Journal of Potato Research, 76, 215–219. Hoy, C. W. & Head, G. (1995). Correlation between behavioral and physiological responses to transgenic potatoes containing Bacillus thuringiensis δ-endotoxin in Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae). Journal of Economic Entomology, 88, 480– 486. Hoy, C. W., Head, G. & Hall, F. R. (1998). Spatial heterogeneity and insect adaptation to toxins. Annual Review of Entomology, 43, 571–594. Iason, G. R. & Villalba, J. J. (2006). Behavioral strategies of mammal herbivores against plant secondary metabolites, the avoidance–tolerance continuum. Journal of Chemical Ecology, 32, 1115–1132. Jallow, M. F. A. & Hoy, C. W. (2005). Phenotypic variation in adult behavioral response and offspring fitness in Plutella xylostella (Lepidoptera: Plutellidae) in response to permethrin. Journal of Economic Entomology, 98, 2195–2202. Jallow, M. & Hoy, C. W. (2006). Quantitative genetics of adult behavioral response and larval physiological tolerance to permethrin in diamondback moth Plutella xylostella (Lepidoptera: Plutellidae). Journal of Economic Entomology, 99, 1388–1395. Jallow, M. & Hoy, C. W. (2007). Indirect selection for increased susceptibility to permethrin in diamondback moth (Plutella xylostella) (Lepidoptera: Plutellidae). Journal of Economic Entomology, 100, 526– 533. Kareiva, P. (1999). Coevolutionary arms races: is victory possible? Proceedings of the National Academy of Sciences of the USA, 96, 8–10. Koricheva, J. (2002). Meta-analysis of sources of variation in fitness costs of plant antiherbivore defenses. Ecology, 83, 176–190.

Lenormand, T. & Raymond M. (1998). Resistance management, the stable zone strategy. Proceedings of the Royal Society of London B, 265, 1985–1990. McCaffery, A. & Nauen, R. (2006). The Insecticide Resistance Action Committee (IRAC), public responsibility and enlightened industrial self-interest. Outlooks on Pest Management, 17, 11–14. Mota-Sanchez, D., Bills, P. S. & Whalon, M. E. (2002). Arthropod resistance to pesticides. In Pesticides in Agriculture and the Environment, ed. W. Wheeler, pp. 241– 272. New York: Marcel Dekker, Mota-Sanchez, D., Hollingworth, R. M., Grafius, E. J. & Moyer, D. D. (2006). Resistance and cross-resistance to neonicotinoid insecticides and spinosad in the Colorado potato beetle, Leptinotarsa decemlineata (Say) (Coleoptera: Chrysomelidae). Pest Management Science, 62, 30–37. Nitao, J. K. (1995). Evolutionary stability of swallowtail adaptations to plant toxins. In Swallowtail Butterflies, their Ecology and Evolution, eds. J. M. Scriber, Y. Tsubaki & R. C. Lederhouse, pp. 39–52. Gainsville, FL: Scientific Publishers Inc. Pittendrigh, B. R. & Gaffney, P. J. (2001). Pesticide resistance: can we make it a renewable resource? Journal of Theoretical Biology, 211, 365–375. Pittendrigh, B. R., Gaffney, P. J., Huesing, J. E. et al. (2004). ‘‘Active’’ refuges can inhibit the evolution of resistance in insects towards transgenic insect-resistant plants. Journal of Theoretical Biology, 231, 461–474. Roush, R. T. (1989). Designing resistance management programs: how can you choose? Pesticide Science, 26, 423–441. Strauss, S. Y., Rudgers, J. A., Lau, J. A. & Irwin, R. E. (2002). Direct and ecological costs of resistance to herbivory. Trends in Ecology and Evolution, 17, 278–285. Tabashnik, B. E. & Croft, B. (1982). Managing pesticide resistance in crop–arthropod complexes: interactions between biological and operational factors. Environmental Entomology, 11, 1137–1144. Thompson, G. D. & Head, G. (2001). Implications of regulating insect resistance management. American Entomologist, 47, 6–10. Urech, P. A., Staub, T. & Voss, G. (1997). Resistance as a concomitant of modern crop protection. Pesticide Science, 51, 227–234.

Chapter 16

Assessing environmental risks of pesticides Paul C. Jepson For synthetic chemicals to pose an environmental risk, a complex sequence of events must unfold that result in a toxic compound reaching a site of action within a susceptible organism with resultant impacts on behavior, fitness or survival. For most chemicals, other than pesticides, toxicity to any organism within the environment, including humans, is an unintended consequence of their use. Pesticides are however among a group of compounds that are synthesized and utilized in such a way that they exhibit direct toxicity to particular organisms, specifically those organisms that humans define as pathogens or pests. Humans compete directly with these organisms for the harvestable yield of crop plants and using poisons to remove them and protect the food and fiber supply has been considered an acceptable, even desirable, activity for centuries, despite evidence that adverse effects on non-target organisms can occur (Devine & Furlong, 2007; Kogan & Jepson, 2007).

16.1 Defining environmental risk Environmental risk is normally expressed as the probability that a defined adverse impact or endpoint will occur within a particular organism as a result of exposure to an environmental stressor (Suter, 1993). Throughout this chapter, I will use the term risk broadly when referring to dif-

ferent kinds of pesticide environmental impact. In some cases, the risks that I refer to could be defined more narrowly as the estimate of probability derived from the product of dose or exposure to the toxin and the susceptibility of the organism to it. This narrow definition becomes problematic, however, when referring to ecological impacts that may unfold in populations and communities generations after the original toxic effect, and away from the location where this impact occurred. This is a complex and difficult field to describe fully, but I have attempted to point out where the conceptual and methodological challenges lie when these arise. Although risk to humans is very much a part of the environmental risk spectrum for pesticides, the role of this chapter is to focus on non-human organisms, and to address ecological risks as an important subset of the problems that pesticides pose. Despite the fact that most of the regulatory procedures surrounding pesticides are designed to address risks to humans, it is important to recognize that the assessment and management of human risks alone will not result in adequate protection of living organisms and the environment. There are a number of reasons why nonhuman organisms may be more sensitive to pesticides than humans, and therefore why ecological risk assessment is a critical component of pesticide regulation and management. These reasons

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



include (adapted to be more relevant to pesticides from Suter, 1993):

r The existence of unique routes of non-human





chemical exposure to pesticides including the use of water as a respiratory medium, oral cleaning of the body, and root uptake by plants. The higher susceptibility of certain non-human species to pesticides because of specific toxicological modes of action (e.g. herbicide toxicity to plants that are relatives of weed species or insecticide toxicity to beneficial insects), the presence of unique biochemical or physiological pathways for effects within the organism (e.g. eggshell synthesis in birds), or higher metabolic rates that increase rates of exposure for a given body size. Specific features of size, structure, distribution or behavior (including feeding) that confer greater rates of exposure and uptake, including total immersion in potentially contaminated environmental media such as soil and water, exploitation of disproportionately contaminated diet sources, or possession of membranes or body parts that are particularly permeable to pesticides. The occurrence of mechanisms of action at higher levels of biological organization that are not relevant to humans, including certain population and community impacts that arise through disruption to food webs (e.g. secondary pest outbreaks or pest resurgence that result from disproportionate toxicity to predatory and parasitic species) or disruption to ecological services, including pollination and nutrient cycling. The close coupling of non-human organisms to their environments which confers greater susceptibility to secondary and indirect effects that impact critical food or habitat resources.

The degree of environmental risk associated with a particular pesticide will be a function of the amount of the active ingredient that is applied, the location and placement of the compound, its partitioning, breakdown and transport in the environment, and its toxicity to the organisms that are exposed to it (van der Werf, 1996). These risks are unfortunately widespread

and often severe because of the nature of pesticide use and the properties of the compounds themselves. Environmental contamination occurs because pesticides are delivered to the sites of treatment by very inefficient processes including spray application, that result in only a tiny fraction of the applied material reaching the intended target organism, and a large proportion of the applied chemical being effectively wasted (Graham-Bryce, 1977; Matthews, 2000). Pesticides may also cause contamination as a result of spillage via accidental release or because of deliberate misuse or abuse. Following release, pesticide physicochemical properties then enable the active ingredients to partition between air, soil, water and the biota (Hornsby et al., 1996; Mackay, 2001) where they must persist long enough to expose and affect the pest, diseases or weeds that they target. This persistence, however, enables pesticides to enter untargeted environmental compartments where their mobility enables them to travel beyond the site of treatment into ground and surface waters (Gilliom et al., 2006) and in some cases to globally distant marine and terrestrial ecosystems (Ueno et al., 2003; Kelly et al., 2007). The characteristics of widespread and inefficient use patterns, accidental spillages, deliberate misuse, environmental partitioning, mobility and persistence result inevitably in the exposure of non-pest species to pesticides and their toxic breakdown products. Given that the toxicological modes of action of pesticides are rarely specific to target pest species alone, non-target organisms that pose no threat to agricultural yields or public health may also be susceptible to them and succumb to toxic impacts as a result of being exposed (McGlaughlin & Mineau, 1995; Nimmo & McEwen, 1998). This chapter explores the assessment of pesticide environmental impact once these compounds are marketed and in commercial use. Although environmental risks are evaluated prior to regulatory approval in a number of jurisdictions internationally, the standard of these assessments varies widely, as does the degree of post-release monitoring and evaluation of the regulatory standards that are set. There are also some regions of the world where little if any


environmental risk assessment is undertaken prior to the approval of pesticides for use, and still others where regulatory oversight of the pesticide market is virtually absent. The pesticide label is intended to provide guidance on whether or not a chemical can be used legally in a certain setting, and if the chemical is used as instructed on the label, there is at least an expectation of a low likelihood of adverse environmental harm. Although the regulatory system that operates prior to marketing can work effectively, there is evidence that environmental impacts do in fact occur even as a result of widespread, legal pesticide application. It is not the role of this chapter to review critically the effectiveness of the pesticide regulatory system: it simply begins from the premise that environmental impacts are associated with pesticide use, no matter how comprehensive the local regulatory system may be, and that assessment and understanding of these impacts will contribute to the adoption of sustainable IPM systems, and the avoidance of environmental harm.

16.2 Evidence for pesticide-related environmental risks Although many literature sources cite the most severe environmental impacts of pesticides, there are very few if any comprehensive summaries of the broader evidence for pesticide impacts on the environment. This gives the impression that the realization of pesticide environmental risks is exceptional and unusual, rather than widespread and normal, which is more probably the case (Devine & Furlong, 2007). The following review provides a critical analysis of the most readily available sources of information concerning pesticide environmental risks. It provides guidance on the assembly of these data for particular chemicals to enable an assessment of the potential for environmental harm in a given set of circumstances. Data that pertain to the exact set of local conditions and potential uses are rarely available, and thus when making these assessments, it is extremely important to draw data from as many sources as possible. It is also important to identify, for each data source, the similarities and differences between the environ-

mental and ecological contexts presented and the context that prompted the assessment in the first place. As a whole, the assembled data set from these disparate sources is more likely to equate with impacts across a large scale than localized impacts.

16.2.1 Pesticide incident monitoring and reporting Wildlife deaths that result from exposure to pesticides are reported in some countries and reveal that incidents involving terrestrial wildlife are still a common occurrence (de Snoo et al., 1999). The majority of incidents are caused by deliberate abuse and misuse of pesticides rather than approved use, but this may reflect the fact that vertebrate cadavers are less concealed and more likely to be detected at frequently visited, open sites, where deliberate abuse occurs, rather than within crop canopies on private land, where legal uses take place. In the UK in 2005, 103 wildlife mortality incidents were attributed to pesticides, the majority of which were caused by organophosphate and carbamate insecticides and various rodenticides (Barnett et al., 2006). Spillages during manufacturing, storage or transport may also provide documentation of ecological hazards, or the risk that these may occur. Rice crop losses in Liaoning Province in China in 1997 provided evidence of herbicide contamination in irrigation water from the Tiaozi and Zhaosutai Rivers that was eventually attributed to accidental spillage upstream (Li et al., 2007). Acts of warfare and looting at a pesticide store in Somalia in 1988 caused contamination by organochlorine and organophoshate pesticides (Lambert, 1997). Five years later, reptiles were absent where surface residues exceeded 10 ppm, they avoided zones contaminated at levels above 1 ppm and reptile species richness was reduced in the valley where the spillage occurred. Additionally surface residues were still lethal to certain reptiles and amphibians and pesticide residue burdens were detectable in ground lizards and in both well and rainpool-inhabiting frogs.

16.2.2 Biological monitoring and surveillance There is a long history of biological monitoring programs in agricultural systems, particularly




in Western Europe, many of which have detected changes, mainly declines, in the abundance and diversity of multiple taxa (Robinson & Sutherland, 2002; Jepson, 2007a). Pesticides may have played a role in these declines, but unless monitoring programs are designed to explicitly address specific mechanisms, they can rarely be used to unambiguously determine cause (Noon, 2003). For example, declines in UK farmland birds have been particularly severe (Ormerod & Watkinson, 2000), but although changes in food quality or quantity on the farm, partly driven by herbicide and insecticide use, have been cited as possible influences (Benton et al., 2002; Newton, 2004), many other factors including climate and land use change have also played a role. A unique and historical example of the use of the biological monitoring data is the retrospective analysis of eggshell thickness in peregrine falcons (Falco peregrinus) from preserved collections in the UK. This analysis revealed the abrupt onset of eggshell thinning in 1947, coinciding with the introduction and widespread use of DDT and its gradual recovery following the banning of this insecticide (Radcliffe, 1967, 1993). More targeted, often shorterterm monitoring investigations may also be used to obtain evidence of pesticide impacts. For example, a 49-site survey by Good & Giller (1991) demonstrated reduced species richness in Staphylinidae (Coleoptera) within fields treated with the organophosphate pesticide dimethoate.

16.2.3 Chemical monitoring and surveillance Environmental monitoring of pesticide residues is more recent, and less widespread than biological monitoring. Contamination of fresh water as a result of runoff, leaching, droplet and vapor drift and over-spraying is still a very serious problem, and large-scale monitoring in the USA revealed that more than half of the streams sampled had pesticide concentrations that exceeded benchmark concentrations for aquatic life (Gilliom et al., 2006). Pesticides also volatilize at the site of application and may move downwind and deposit at cooler locations a significant distance away from the site of application. Monitoring of summertime atmospheric transport of non-persistent insecticides, fungicides and herbicides in California’s Central Valley has shown that air and surface

water concentrations in the Valley, and the Sequoia National Park which lies downwind, are directly proportional to levels of use of these compounds (LeNoir et al., 1999). Pesticides were also detected in the rain and snow precipitating in the same mountain range (McConnell et al., 1998). Global monitoring of persistent organochlorine pesticides in skipjack tuna (Katsuwonus pelamis) tissue has found residues in all of the fish sampled from a number of oceanic systems (Ueno et al., 2003), and these compounds have also been shown to accumulate in terrestrial food webs, including birds and mammals in the Arctic (Kelly et al., 2007). Ecotoxicologists are still developing methods to understand the effects of complex pesticide mixtures (e.g. Posthuma et al., 2002), and our understanding of the possible effects of long-term, multiple, low-dose exposures is still very limited. Despite these constraints, chemical monitoring data are vital in enabling mapping of contamination, demonstration of the potential for exposure and identification of compounds of potential concern for risk assessment and risk management purposes.

16.2.4 Risk estimates derived from chemical monitoring Monitoring data can only be translated into estimates of ecological impacts if they are supported by measurements of effects or by risk assessment procedures that relate chemical exposure estimates to specific biological endpoints. Streams and drainage canals in the Yakima River Basin in Washington State, USA with the highest and most potentially toxic pesticide concentrations tended to have the highest numbers of pollution-tolerant benthic aquatic invertebrates (Fuhrer et al., 2004), implying that more sensitive species have been lost as a result of pesticide toxicity. Organophosphorous residues and depressed cholinesterase activity in amphibian tissues from the California Sierra Nevada, USA, were associated with sites where there was poor to moderate amphibian population status (Sparling et al., 2001). The inferential power inherent within the chemical monitoring data may however be poor, and it is not always possible to assert causality if there are no control comparisons, or if there are other stressors that confound the data.


16.2.5 Published field experiments There is a limited literature from field experiments undertaken in open natural or agricultural systems that contains evidence concerning the impacts of individual and multiple pesticides on a wide array of non-target taxa. At best, properly designed experiments engender a high degree of statistical and inferential power. With the rigorous statement of hypotheses, effective and realistic design combined with careful pesticide application procedures and efficient sampling of organisms, they may provide powerful evidence of the nature, level and duration of pesticide impacts. Experiments in open freshwater ecosystems Although investigations made in natural still (lentic) and flowing (lotic) surface waters may not be laid out as formal experiments, they may carry a high degree of inferential power if they adhere to rigorous procedures for design, layout and analysis. There are however some substantial methodological challenges. It may be difficult to quantify the chemical exposure of the stream flora and fauna because it is often not possible to account for and measure all the sources of pesticide contamination, and because invertebrates and particularly fish are highly mobile in lotic systems. Without mechanistic hypotheses concerning potential impacts, it may also be impossible to define the most appropriate ecological endpoints to measure, and many studies simply record abundance and diversity of readily sampled taxa. The sites themselves may not lend themselves to selection of control or reference areas that are unaffected by pesticide exposure, and matched sites that are otherwise identical in terms of hydrology, riparian habitat characteristics, pond and stream chemical and physical conditions; other potential stressors and surrounding land uses may not be available. Leonard et al. (2000) investigated storm runoff of the organochlorine insecticide endosulfan that could potentially lead to fish kills in the Namoi River, Australia. They sought evidence for impacts in macroinvertebrate communities because these occurred in high densities and were relatively sedentary, compared with fish. Population densities were reduced at downstream sites with 10– 25-fold higher pesticide concentrations compared

with multiple reference sites in upstream reaches. The weight of the evidence from this study pointed to endosulfan exposure being responsible for these effects. The authors were able to make this assertion because they employed a relatively symmetrical BACI (Before–After–Control– Impact) sampling design that included several control sites, thereby avoiding the low statistical and inferential power associated with asymmetrical studies that include only a single reference or control location (Underwood, 1994). They also had mechanistic evidence for pesticide effects on key taxa that were depleted in the contaminated sites because they had previously demonstrated that riverine pesticide concentrations exceeded the 48-h LC50 s for mayfly (Ephemeroptera) and caddisfly (Trichoptera) nymphs from the same location (Leonard et al., 1999). Thiere & Schulz (2004) found large differences in both turbidity and pesticide exposure between agricultural and upstream reaches of the Lourens River in South Africa. They argued that either or both of these may be responsible for differences in macroinvertebrate community structure between the two sites that left only the more waterquality-insensitive taxa in agricultural reaches. The requirements for high statistical and inferential power are far better satisfied when treatments can be allocated to separate, matched replicate aquatic habitats at a range of treatment concentrations; these conditions occur in the case of temporary pond invertebrate communities exposed to pesticide drift during locust control operations in the Sahelian zone of West Africa. Lahr et al. (2000) detected negative direct (toxicological) effects of fenitrothion, diflubenzuron, deltamethrin and bendiocarb on cladocerans, fairy shrimps and backswimmers, and also found indirect effects, particularly patterns of superabundance, that could be explained by reduced predation or competition in treated ponds. Only 12 of the 80 taxa in the ponds that they studied, however, satisfied the requirements for statistical analysis, and the fluctuating populations and low abundance of these taxa meant that inferences about effects were drawn from a combination of statistics, graphical analysis and expert judgment, combined with evidence for direct toxicity derived from laboratory experiments (Lahr et al., 2001).




Experiments in open freshwater agroecosystems Some of the most sophisticated understanding of pesticide impacts in aquatic systems have come from IPM studies in rice agriculture, as opposed to ecotoxicological investigations in natural water bodies. Combined observational and experimental studies in Indonesian rice demonstrated for example, that carbofuran and monocrotophos applications caused pest resurgence by suppressing predators and their alternative food supplies early in the season (Settle et al., 1996). This investigation, and others in rice, are among the only research studies of pesticide impacts to address the importance of trophic linkages and their disruption in sprayed systems; in this case the linkages among organic matter, detritivores, plankton feeders and generalist predators that result in herbivore (pest) suppression early in the season by generalist natural enemies that do not rely solely upon pest herbivores for their survival. Deltamethrin applications to rice have been found to increase rice herbivores (mainly Delphacidae) by 4 million/ha per sampling date, and decrease natural enemies by 1 million/ha per sampling date, significantly reducing plant-borne invertebrate food web length through the removal of certain spider, coccinellid and heteropteran predators (Schoenly et al., 1994). The result of this food web simplification and associated disorganization of the natural enemy community is greatly increased unpredictability in pest population dynamics (Cohen et al., 1994). Experiments in open fields and farms: impacts on invertebrates Terrestrial agriculture is very poorly supported by investigations of pesticide impacts on agriculturally relevant scales. Investigators tend to be lured by the simplicity and logistic tractability of replicated small plot experiments and lose sight of the fact that the small scale of these investigations is trivial in comparison with the scale upon which the population processes of non-target organisms ensue (Jepson, 1989, 1993, 2007b); hence, effects may be routinely unrecorded or underestimated. Perhaps because they are not confined solely to an agricultural matrix, ecotoxicological investigations of the side effects of disease vector and

plague pest pesticides tend to be undertaken on scales that are ecologically relevant. Malaise traps and fallout funnels beneath sprayed trees captured invertebrates and demonstrated reductions in Diptera and Hymenoptera after aerial and ground-based endosulfan treatments against tsetse fly (Glossina spp.) (Everts et al., 1983). All pool-dwelling fish were also killed. Pitfall traps and Malaise traps in 3–15-ha tsetse fly treatments with synthetic pyrethroids also exhibited reductions in multiple invertebrate taxa, including a ctenizid spider that was selected as an indicator taxon (Everts et al., 1985). Barrier treatments over 5–20-km2 areas against locusts using the insect growth regulator diflubenzuron enabled sweep net comparisons between invertebrates in 50-mwide sprayed and 600-m-wide unsprayed strips (Tingle, 1996). Assessments were made of 300 species, distributed between 120 families and 17 orders, and revealed significant reductions in Lepidoptera and Acrididae. Investigations on this scale enable the potential of unsprayed areas as refuges for non-target taxa to be determined, and also the ecological selectivity of barrier spraying, which biases exposure towards mobile animals that walk through the treated area and intercept sprayed strips (Jepson & Sherratt, 1996). Experimental regimes may also be undertaken on a large scale, but there are trade-offs to address between replication and plot size. Peveling et al. (1999a) argued that replicated 16-ha plots were too small to enable assessment of effects on flying invertebrates, but larger 400ha plots could not be replicated in their investigation. They demonstrated >75% reductions for more than 3 months in Collembola, Formicidae, Carabidae and Tenebrionidae in fenitrothiontreated plots. They found successively more limited impacts with a fenitrothion–esfenvalerate mixture and with the insect growth regulator triflumuron. Impacts of triflumuron on non-target Lepidoptera may, however, have been among the most significant and damaging effects detected, because 70% of these species were endemic to the area of Madagascar where this investigation took place. Peveling et al. (1999b) also explored the use of presence/absence assessments in replicated, intermediate-sized plots (50 ha) in order to develop a more rapid risk assessment program.


They again demonstrated that fenitrothion had significant impacts on non-target invertebrates, with 75% of Carabidae, Tenebrionidae, Formicidae and Ephydridae reduced significantly in the pesticide treatment. In temperate systems, one of the only investigations to address realism in spatial and temporal scales is the Boxworth Study, which investigated the impacts of different spray regimes allocated to contiguous blocks of whole fields in the UK (Greig-Smith et al., 1992). When whole fields were treated with organophosphate insecticides in a high input regime, certain ground beetles (Carabidae) became locally extinct (Burn, 1992). Jepson & Thacker (1990) demonstrated experimentally that Carabidae reinvade organophosphate pesticide-treated areas as a result of random dispersal behavior, once residual toxicity has ameliorated after 7–9 days (Unal & Jepson, 1991), but that this process can take up to 30 days at distances of 100 m into treated plots. Within-field, replicated experiments on unrealistically small scales will therefore not detect significant impacts on dispersive taxa because these will invade rapidly from untreated control areas. Modeling (Sherratt & Jepson, 1993) demonstrated that delayed reinvasion by Carabidae of the larger, organophosphatetreated whole fields in sprayed systems could result in the local extinctions detected by Burn (1992), and they also predicted that these effects could result in locally increased pest density (also termed pest resurgence). This was subsequently confirmed experimentally by Duffield et al. (1996), who demonstrated a wave of aphid pest and collembolan population outbreaks in advance of the reinvasion front of epigeal predatory species that were temporarily impacted by spray applications. Large or realistically scaled experiments tend to reveal the long-term effects of broad-spectrum insecticides that are not detected within smallerscale investigations. Our overall understanding of the effects of broad-spectrum insecticides in agroecosystems might be transformed if experiments were conducted on scales that equated better with the scales of commercial treatment and the scales over which the population processes of non-target organisms takes place. At risk however is the low statistical power inherent in

experiments with low levels of replication. Experimental research into the ecological impacts of GM crops expressing insecticidal genes compared with conventional insecticides has demonstrated that smaller, replicated plots may offer an effective alternative to large plots in certain commodities such as cotton, and that statistical power may be further strengthened by conducting multi-year investigations (Naranjo, 2005). Prasifka et al. (2005) recommended that small plots should be avoided altogether for investigating pesticide and GM crop ecological impacts in corn, and that intermediatesized plots could be isolated from one another by cultivated strips to enable effect size to be more accurately measured. Experiments in open fields and farms: impacts on vertebrates Assessments of pesticide impacts on vertebrate wildlife also comprise a relatively limited international literature, with impacts on birds being the best represented. Plague pest and vector management campaigns in Africa may constitute some of the largest scale wildlife exposures that occur, but data concerning impacts on lizards, birds and small mammals are relatively sparse. A non-replicated set of 800-ha treatments, with paired controls, revealed that a low dosage of a fenitrothion–esfenvalerate mixture significantly reduced the abundance of the endemic lizard Chalarodon madagascarensis by 39%, possibly through direct mortality of juvenile lizards (Peveling & Nagel, 2001). There were similar but nonsignificant reductions in a triflumuron treatment. The same lizard species, with another species, Mabuya elegans, and the lesser hedgehog tenrec (Echinops telfairi), are termite predators and the potential for food chain perturbations caused by pesticides was demonstrated when all three species were reduced following locust control barrier treatments over 45 km2 with fipronil that caused large reductions in the harvester termite Coarctotermes clepsydra (Peveling et al., 2003). This compound has also been shown, however, to be directly toxic to lizards and both direct and indirect effects may underlie observed effects (Peveling & Demba, 2003). Mineau (2002) reviewed 181 field investigations of cholinesterase-inhibiting pesticide effects




on birds in the literature, encompassing field and pasture crops, forestry, orchards, mosquito control, forage crop or turf treatments and exposure via drinking (i.e. from puddles, irrigation equipment, or crop leaf whorls). Of these, no effects were detected in 67 studies, sublethal effects only were found in 23, lethal effects were detected in 60, and 31 exhibited mass bird mortality. Mineau also demonstrated by probabilistic modeling that pesticide application rate, intrinsic toxicity, relative dermal toxicity and possibly volatility were all critical factors in explaining the effects exhibited in these investigations. These models contributed to predictions of the widespread lethal risk of pesticides to birds from insecticide use in the USA (Mineau & Whiteside, 2006). This analysis revealed that the crops most susceptible to bird mortality in the USA were corn and cotton.

16.2.6 Published laboratory and outdoor enclosed-system experiments Laboratory bioassay data may provide essential mechanistic underpinning to field investigations and evidence for potential environmental risks. They certainly have a degree of explanatory power, and lie at the heart of the complex risk assessment procedures employed by regulatory agencies. The bioassay data of greatest explanatory power are those that reveal no effects in a particular group of organisms, at least in the concentration ranges encountered in the field. If a pesticide is simply not toxic, then mechanisms of toxicity can not be recruited in the analysis of any observed effects in the field. When toxicity is revealed, the likelihood that this will be translated into actual environmental impacts, and the relationship between these possible impacts and the relative toxicities of different compounds is far more complex and challenging. Evidence of toxicity from laboratorybased bioassays therefore provides evidence of the potential for harm, and provides the justification for further inquiry and analysis. This section critically reviews data and analyses from laboratory and highly controlled field enclosure experiments to provide a basis for interpreting the value of this information in the evaluation of environmental risks. Aquatic toxicologists use tanks, artificial pools or ditches, termed microcosms and larger

mesocosms, to simulate aquatic habitats. These are amenable to experimental manipulation and are thought to offer more realism than laboratorybased, single-species bioassays. The systems used are enormously diverse, as are the flora and fauna selected and the endpoints that are recorded. The most frequently studied taxa are phytoplankton and zooplankton, macrophytes (vascular plants and loose filamentous algae), macroinvertebrates and fish (Brock & Buddle, 1994). The chosen endpoints are also enormously diverse, and include community composition, chlorophyll a content, biomass, abundance, similarity, diversity and spatial distribution. Results reveal primary effects, which many argue are likely to equate to those that may occur in the real world, and secondary effects and recovery dynamics which are also affected by the construction and characteristics of the experimental system. In an indoor system, designed to simulate drainage ditches, Brock et al. (1992a) demonstrated effects of chlorpyrifos on amphipods, insects and isopods, and also revealed the importance of macrophytes in determining water and sediment concentrations of the pesticide. Secondary effects in the same system were also affected by the presence of macrophytes, and included trophic effects on primary producers, herbivores and other functional groups, that resulted from loss of arthropods (Brock et al., 1992b). Functional impacts were also detected, particularly effects on community metabolism, signaled by decreases in dissolved oxygen and pH, and increases in alkalinity and conductivity (Brock et al., 1993). There is a large and critical literature concerning the degree to which data from mesocosms can contribute to regulatory toxicology (e.g. Shaw & Kennedy, 1996; Maund et al., 1997). Artificial pond systems may not relate closely to any natural system (e.g. Williams et al., 2002); however, a recent review has found consistent patterns of response, relative to laboratory-based single-species toxicity tests which revealed that the ecological effects threshold is normally about ten times the maximum concentrations that are determined to be acceptable in the laboratory tests (van Wijngaarden et al., 2005). The use of laboratory-based bioassay data in the evaluation of pesticide risks to non-target invertebrates can be equally problematic, although these


data are available for hundreds of compounds and thousands of invertebrate species (e.g. Croft, 1990). In the case of these organisms so-called semi-field data, from artificial enclosures, although useful, are only available for a small number of pesticides and taxa (Jepson, 1993). Stark et al. (1995) used several methods for extrapolating laboratory test data to predict field effects against a number of predatory and parasitic invertebrates and found that these gave variable predictions of the compatibility of these compounds with IPM. Although laboratory test data are widely used to compare pesticides and determine their compatibility with IPM, in reality, these data must be considered alongside information concerning exposure rate, chemical fate and invertebrate behavior if they are to contribute to predictions of either short-term or long-term ecological effects (Jepson, 1989, 1993, 2007b). Differences in key life history variables determine the population trajectories of affected populations after initial toxic effects have taken place, further questioning the basis for using toxicological data alone in the evaluation of the potential risks that compounds pose (Stark et al., 2004). One way of overcoming the requirement for large plot sizes to compensate for the dispersed populations and mobility of small mammals, is to enclose small populations within replicated barriers. Edge et al. (1996) demonstrated a dose-response by vole (Microtus canicaudus) populations in a series of replicated enclosures, when they were exposed to a single treatment with azinphos-methyl. They argued that the short-term nature of the impact could reflect a rapid population response by the voles, and that effects might be severe when voles were exposed to multiple treatments. Enclosure techniques in this case, may have presented the only effective method for obtaining data concerning pesticide side effects.

16.3 Conclusions This chapter has outlined the sources of data and approaches that are used for the evaluation of pesticide environmental risks. It has argued implicitly that evidence for potential effects can be drawn from a number of sources, relying most

heavily on those data that reveal impacts under conditions of field use. This is an evidence-based approach, which requires critical evaluation of the data derived from each of the sources that are listed. Table 16.1 provides a summary of the main constraints and uses associated with data derived from the sources of evidence for pesticide environmental impacts outlined in this chapter. In a particular location, data sets from these sources may be assembled for specific compounds or closely related materials, but these must be qualified by an assessment of their relevance to local conditions. These data may also be used to aid interpretation of observations or experiments, and also to assist end-users in ranking pesticides for their potential to inflict environmental harm. Some materials may accumulate a positive “environmental profile” because for example, chemical monitoring for these specific materials does not detect them, and because widespread use does not generate evidence for potential impacts from either monitoring data or experiments. Other materials may warrant far greater caution, because they have been repeatedly detected in concentrations that may cause adverse effects, and also because well-designed monitoring programs and experiments with these compounds have sufficient inferential power to attribute cause. In summarizing some of the key challenges associated with some of these data sources, the chapter has revealed the conceptual challenges faced by organizations and individuals that might wish to develop more quantitative risk assessments. The chapter has not referred to simple environmental impact indices that are advanced by some as useful tools in IPM decision making or evaluation, because in the author’s opinion, none of these to date has been adequately validated for use in real-world decision making. Similarly, it has not referred to the complex risk assessment procedures used by regulatory agencies, which must undertake this process for hundreds of compounds and use data sets that are usually built from laboratory bioassays and chemical fate and behavior data, prior to extensive field use. The author argues that there is a role for the characterization of compounds, based upon the properties


Table 16.1 Summary of the main constraints and uses associated with data derived from the sources of evidence for pesticide environmental impacts outlined in this chapter Source of evidence Pesticide incident reporting

Biological monitoring and surveillance

Chemical monitoring and surveillance

Risk estimates from chemical monitoring

Published field experiments

Published laboratory and outdoor, enclosed system experiments

Major constraints

Main uses for data


• Evaluation of status and trends in Only a limited number of incident wildlife poisoning as contributory reporting schemes internationally, factor to population status of and requires high quality forensic threatened or endangered species and chemical analysis facilities, • Early warning of unexpected hazards combined with an alert public; associated with specific compounds may be aggregated to a national in commercial use or international scale • Detection and prosecution of illegal use and abuse, and monitoring of compliance with the law • Normally not possible to • Evaluation of status and trends in Monitoring and surveillance are definitively attribute cause, or to wildlife populations associated with rarely sensitive enough to detect tease out the contribution of agriculture anthropogenic changes in floral or pesticide mortality to population faunal biodiversity, or specific status effects associated with pesticides; • Not effective in early warning successful programs are built upon clearly stated hypotheses and rigorous design, including control areas Expensive, logistically and • Measurement of off-crop movement When available, these data are methodologically complex, and and environmental distribution of extremely informative and helpful temporal and spatial resolution pesticides in developing risk assessments; normally poor • Contributes to accurate when sufficiently localized and environmental concentration detailed, they can be used to estimates for risk assessment backtrack to particular uses that may then be addressed by IPM programs • Very effective early warning tool for potential ecological effects • Evaluation of the benefits of IPM programs and regulatory instruments Risk estimates based upon • Mapping and identification of places Most effective when follow-up toxicological data are not readily where impacts may be unacceptably assessments are possible in situ, extrapolated to the field, even high following risk assessments that when based upon accurate • Support for regulatory instruments, identify the possibility that chemical monitoring; the including the setting of limits adverse impacts may be occurring ecological context and ecological • Very effective in developing early processes must also be taken into warning for potential ecological consideration effects from chemical monitoring data • Evaluation of the benefits of IPM programs and regulatory instruments, particularly when followed up with field measurements Highly context specific, and difficult • Testing of hypotheses that address Data from multiple experiments may to extrapolate to larger scales and specific mechanisms of be assembled within databases, or other locations until a body of environmental harm employed within meta-analysis of complementary experimental • Development of reduced-risk IPM environmental impacts evidence has been assembled regimes • Verification or further ecological exploration of initial risk assessments Exposure pathways and pesticide • Essential data for the derivation of Toxicological datasets are extremely concentrations may not be risk estimates when combined with useful when handled and sufficiently relevant to field environmental concentrations interpreted appropriately; conditions for simple assessments • Evidence of low, or no toxicity for particularly data that demonstrate of environmental risk certain materials is of great value in no effects IPM program design Biased towards cases where wildlife cadavers are readily detectable



and effects that they reveal in widespread field use, where chemicals confront ecological communities and processes and exert impacts that we must respond to and manage. IPM researchers and practitioners should not be inhibited or discouraged from assembling these data sets and discussing them with stakeholders as an additional element in the complex decision support process that underlies IPM.

References Barnett, E. A., Fletcher, M. R., Hunter, K. & Sharp, E. A. (2006). Pesticide Poisoning of Animals in 2005. London: Department of Environment and Rural Affairs, Pesticide Safety Division. Available at PSD/finalPPA2005report.pdf Benton, T. G., Bryant, D. M., Cole, L. & Crick, H. Q. P. (2002). Linking agricultural practices to insect and bird populations: a historical study over three decades. Journal of Applied Ecology, 39, 673–687. Brock, T. C. M. & Buddle, B. J. (1994). On the choice of structural parameters and endpoints to indicate responses of freshwater ecosystems to pesticide stress. In Freshwater Field Tests for Hazard Assessment of Chemicals, eds. I. R. Hill, F. Heimbach, P. Leeuwangh & P. Matthiessen, pp. 19–56. Boca Raton, FL: CRC Press. Brock, T. C. M., Crum, S. J. H., van Wijngaarden, R. et al. (1992a). Fate and effects of the insecticide Dursban 4E in indoor Elodea-dominated and macrophyte-free freshwater model ecosystems. I. Fate and primary effects of the active ingredient chlorpyrifos. Archives of Environmental Contamination and Toxicology, 23, 69–84. Brock, T. C. M., van den Bogaert, M., Bos, A.R. et al. (1992b). Fate and effects of the insecticide Dursban 4E in indoor Elodea-dominated and macrophyte-free freshwater model ecosystems. II. Secondary effects on community structure. Archives of Environmental Contamination and Toxicology, 23, 391–409. Brock, T. C. M., Vet, J. J. R. M., Kerkhofs, M. J. J. et al. (1993). Fate and effects of the insecticide Dursban 4E in indoor Elodea-dominated and macrophyte-free freshwater model ecosystems. III. Aspects of ecosystem functioning. Archives of Environmental Contamination and Toxicology, 25, 160–169. Burn, A. J. (1992). Interactions between cereal pests and their predators and parasites. In Pesticides and the Environment: The Boxworth Study, eds. P. Greig-Smith, G. H. Frampton & A. Hardy, pp. 110–131. London: Her Majesty’s Stationery Office (HMSO).

Cohen, J. E., Schoenly, K., Heong, K. L. et al. (1994). A food web approach to evaluating the effect of insecticide spraying on insect pest population dynamics in a Philippine irrigated rice ecosystem. Journal of Applied Ecology, 31, 747–763. Croft, B. A. (1990). Arthropod Biological Control Agents and Pesticides. New York: Wiley Interscience. de Snoo, G. R., Scheidegger, N. M. I. & de Jong, F. M. W. (1999). Vertebrate wildlife incidents with pesticides: a European survey. Pesticide Science, 55, 47–54. Edge, W. D., Carey, R. L., Wolff, J. O., Ganio, L. M. & Manning, T. (1996). Effects of Guthion 2S on Microtus canicaudus: a risk assessment validation. Journal of Applied Ecology, 33, 269–278. Everts, J. W., Van Frankenhuyzen, K., Roman, B. et al. (1983). Observations on side effects of endosulfan used to control tsetse in a settlement area in connection with a campaign against human sleeping sickness in Ivory Coast. Tropical Pest Management, 29, 177–182. Everts, J. W., Kortenhoff, B. A., Hoogland, H. et al. (1985). Effects on non-target terrestrial arthropods of synthetic pyrethroids used for the control of tsetse fly (Glossina spp.) in settlement areas of the southern Ivory Coast, Africa. Archives of Environmental Contamination and Toxicology, 14, 641–650. Devine, D. J. & Furlong, M. J. (2007). Insecticide use: contexts and ecological consequences. Agriculture and Human Values, 24, 281–306. Duffield, S. J., Jepson, P. C., Wratten, S. D. & Sotherton, N. W. (1996). Spatial changes in invertebrate predation pressure in winter wheat following treatment with dimethoate. Entomologia Experimentalis et Applicata, 78, 9–17. Fuhrer, G. J., Morace, J. L., Johnson, H. M. et al. (2004). Water Quality in the Yakima River Basin, Washington, 1999– 2000, US Geological Survey Circular No. 1237. Reston, VA: US Department of the Interior. Gilliom, R. J., Barbash, J. E., Crawford, C. G. et al. (2006). The Quality of our Nation’s Waters: Pesticides in the Nation’s Streams and Ground Water, 1992–2001, US Geological Survey Circular No. 1291. Reston, VA: US Department of the Interior. Good, J. A. & Giller, P. S. (1991). The effect of cereal and grass management on staphylinid (Coleoptera) assemblages in South-East Ireland. Journal of Applied Ecology, 28, 810–826. Graham-Bryce, I. J. (1977). Crop protection: a consideration of the effectiveness and disadvantages of current methods and of the scope for improvement. Philosophical Transactions of the Royal Society of London B, 281, 163–179.


Greig-Smith, P., Frampton, G. H. & Hardy, A. (eds.) (1992). Pesticides and the Environment: The Boxworth Study. London: Her Majesty’s Stationery Office (HMSO). Hornsby, A. G., Wauchope, R. D. & Herner, A. E. (1996). Pesticide Properties in the Environment. New York: SpringerVerlag. Jepson, P. C. (ed.) (1989). Pesticides and Non-Target Invertebrates. Wimborne, UK: Intercept. Jepson, P. C. (1993). Insects, spiders and mites. In Handbook of Ecotoxicology, ed. P. Calow, pp. 299–325. Oxford, UK: Blackwell Scientific Publications. Jepson, P. C. (2007a). Challenges to the design and implementation of effective monitoring for GM crop impacts: lessons from conventional agriculture. In Genetically Modified Organisms in Crop Production and their Effects on the Environment: Methodologies for Monitoring and the Way Ahead, eds. K. Ghosh & P. C. Jepson, pp. 34– 57. Rome, Italy: Food and Agriculture Organization of the United Nations. Jepson, P. C. (2007b). Ecotoxicology and IPM. In Perspectives in Ecological Theory and Integrated Pest Management, eds. M. Kogan & P. C. Jepson, pp. 522–551. Cambridge, UK: Cambridge University Press. Jepson, P. C. & Sherratt, T. N. (1996). The dimensions of space and time in the assessment of ecotoxicological risks. In ECOtoxicology: Ecological Dimensions, eds. D. J. Baird, P. W. Grieg-Smith & P. E. T. Douben, pp. 44–54. London: Chapman & Hall. Jepson, P. C. & Thacker, J. R. M. (1990). Analysis of the spatial component of pesticide side-effects on non-target invertebrate populations and its relevance to hazard analysis. Functional Ecology, 4, 349–358. Kelly, B. C., Ikonomou, M. G., Blair, J. D., Morin, A. E. & Gobas, F. A. P. C. (2007). Food-web specific biomagnification of persistent organic pollutants. Science, 317, 236–239. Kogan, M. & Jepson, P. C. (2007). Ecology, sustainable development and IPM. In Perspectives in Ecological Theory and Integrated Pest Management, eds. M. Kogan & P. C. Jepson, pp. 1–44. Cambridge, UK: Cambridge University Press. Lahr, J., Diallo, A. O., Gadji, B. et al. (2000). Ecological effects of experimental insecticide applications on invertebrates in Sahelian temporary ponds. Environmental Toxicology and Chemistry, 19, 1278–1289. Lahr, J., Badji, A., Marquenie, S. et al. (2001). Acute toxicity of locust insecticides to two indigenous invertebrates from Sahelian ponds. Ecotoxicology and Environmental Safety, 48, 66–75. Lambert, M. R. K. (1997). Environmental effects of heavy spillage from a destroyed pesticide store near

Hargeisa (Somaliland) assessed during the dry season, using reptiles and amphibians as bioindicators. Archives of Environmental Contamination and Toxicology, 32, 80–93. LeNoir, J. S., McConnell, L. L., Fellers, G. M., Cahill, T. M. & Seiber, J. N. (1999). Summertime transport of current-use pesticides from California’s Central Valley to the Sierra Nevada mountain range, USA. Environmental Toxicology and Chemistry, 18, 2715–2722. Leonard, A. W., Hyne, R. V., Lim, R. P. & Chapman, J. C. (1999). Effect of endosulfan runoff from cotton fields on macroinvertebrates in the Namoi River. Ecotoxicology and Environmental Safety, 42, 125–134. Leonard, A. W., Hyne, R. V., Lim, R. P., Pablo, F. & van den Brink, P. J. (2000). Riverine endosulfan concentrations in the Namoi River, Australia: link to cotton field runoff and macroinvertabrate population densities. Environmental Toxicology and Chemistry, 19, 1540–1551. Li, Q., Luo, Y., Song, J. & Wu, L (2007). Risk assessment of atrazine polluted farmland and drinking water: a case study. Bulletin of Environmental Contamination and Toxicology, 78, 187–190. Maund, S. J., Sherratt, T. N., Strickland, T. et al. (1997). Ecological considerations in pesticide risk assessment for aquatic systems. Pesticide Science, 49, 185–190. McConnell, L. L., LeNoir, J. S., Datta, S. & Seiber, J. N. (1998). Wet deposition of current use pesticides in the Sierra Nevada mountain range, California, USA. Environmental Toxicology and Chemistry, 17, 1908–1916. McGlaughlin, A. & Mineau, P (1995). The impact of agricultural practices on biodiversity. Agriculture, Ecosystems and the Environment, 55, 201–212. Mackay, D. (2001). Multimedia Environmental Models: The Fugacity Approach. Boca Raton, FL: CRC Press. Matthews, G. A. (2000). Pesticide Application Methods. Oxford, UK: Blackwell Science. Mineau, P. (2002). Estimating the probability of bird mortality from pesticide sprays on the basis of the field study record. Environmental Toxicology and Chemistry, 21, 1497–1506. Mineau, P. & Whiteside, M. (2006). Lethal risk to birds from insecticide use in the United States: a spatial and temporal analysis. Environmental Toxicology and Chemistry, 25, 1214–1222. Naranjo, S. E. (2005). Long-term assessment of the effects of transgenic Bt cotton on the abundance of nontarget arthropod natural enemies. Environmental Entomology, 34, 1193–1210. Newton, I. (2004). The recent declines of farmland bird populations in Britain: an appraisal of causal factors and conservation actions. Ibis, 146, 579–600.




Nimmo, D. R. & McEwen, L. C. (1998). Pesticides. In Handbook of Ecotoxicology, vol. 2, ed. P. Calow, pp. 619–667. Oxford, UK: Blackwell Scientific Publications. Noon, B. R. (2003). Conceptual issues in monitoring ecological resources. In Monitoring Ecosystems: Interdisciplinary Approaches for Evaluating Ecoregional Initiatives, eds. D. E. Busch & J. C. Trexler, pp. 27–72. Washington, DC: Island Press. Ormerod, S. J. & Watkinson, A. R. (2000). Special profile: birds and agriculture. Journal of Applied Ecology, 37, 699–705. Peveling, R. & Demba, S. A. (2003). Toxicity and pathogenicity of Metarhizium anisopliae var. acridium (Deuteromycotina, Hyphomycetes) and fipronil to the fringe-toed lizard Acathodactylus dumerili (Squamata: Lacertidae). Environmental Toxicology and Chemistry, 22, 1437–1447. Peveling, R. & Nagel, P. (2001). Locust and Tsetse fly control in Africa: does wildlife pay the bill for animal health and food security? In Pesticides and Wildlife, ed. J. J. Johnston, pp. 82–108. Washington, DC: American Chemical Society. Peveling, R., Rafanomezantsoa, J.-J., Razafinirina, R., Tovonkery, R. & Zafimaniry, G. (1999a). Environmental impact of the locust control agents fenitrothion, fenitrothion–esfenvalerate and triflumuron on terrestrial arthropods in Madagascar. Crop Protection, 18, 659–676. Peveling, R., Attignon, S., Langewald, J. & Oouambama, Z. (1999b). An assessment of the impact of biological and chemical grasshopper control agents on ground-dwelling arthropods in Niger, based on presence/absence sampling. Crop Protection, 18, 323–339. Peveling, R., McWilliam, A. N., Nagel, P. et al. (2003). Impact of locust control on harvester termites and endemic vertebrate predators in Madagascar. Journal of Applied Ecology, 40, 729–741. Posthuma, L., Suter, G. W. & Traas, T. P. (2002). Species Sensitivity Distributions in Ecotoxicology. Boca Raton, FL: Lewis Publishers. Prasifka, J. R., Hellmich, R. L., Dively, G. P. & Lewis, L. C. (2005). Assessing the effects of pest management on nontarget arthropods: the influence of plot size and isolation. Environmental Entomology, 34, 1181–1192. Radcliffe, D. A. (1967). Decrease in eggshell weight in certain birds of prey. Nature, 215, 208–210. Radcliffe, D. A. (1993). The Peregrine Falcon, 2nd edn. London: Poyser. Robinson, R. A. & Sutherland, W. J. (2002). Post-war changes in arable farming and biodiversity in Great Britain. Journal of Applied Ecology, 39, 157–176.

Schoenly, K. G., Cohen, J. E., Heong, K. L. et al. (1994). Quantifying the impact of insecticides on food web structure of rice-arthropod populations in a Philippine farmer’s irrigated field: a case study. In Food Webs: Integration of Patterns and Dynamics, eds. G. A. Polis, G. A. & K. Wisemiller, pp. 343–351. London: Chapman & Hall. Shaw, J. L. & Kennedy, J. H. (1996). The use of aquatic field mesocosm studies in risk assessment. Environmental Toxicology and Chemistry, 15, 605–607. Settle, W. H., Ariawan, H., Astuti, E. T. et al. (1996). Managing tropical rice pests through conservation of generalist natural enemies and alternative prey. Ecology, 77, 1975–1988. Sherratt, T. N. & Jepson, P. C. (1993). A metapopulation approach to modelling the long-term impact of pesticides on invertebrates. Journal of Applied Ecology, 30, 696–705. Sparling, D. W., Feller, G. M. & McConnell, L. L. (2001). Pesticides and amphibian population declines in California, USA. Environmental Toxicology and Chemistry, 20, 1591–1595. Stark, J. D., Jepson, P. C. & Mayer, D. F. (1995). Limitations to use of topical toxicity data for predictions of pesticide side-effects in the field. Journal of Economic Entomology, 88, 1081–1088. Stark, J. D., Banks, J. E. & Vargas, R. (2004). How risky is risk assessment: the role that life history strategies play in susceptibility of species to stress. Proceedings of the National Academy of Sciences of the USA, 101, 732–736. Suter, G. W. (1993). Ecological Risk Assessment. Albany, GA: Lewis Publishers. Thiere, G. & Schulz, R. (2004). Runoff-related agricultural impact in relation to macroinvertebrate communities of the Lourens River, South Africa. Water Research, 38, 3092–3102. Tingle, C. C. D. (1996). Sprayed barriers of diflubenzuron for control of the migratory locust (Locusta migratoria capito (Sauss.)) [Orthoptera: Acrididae] in Madagascar: short-term impact on relative abundance of terrestrial non-target invertebrates. Crop Protection, 15, 579–592. Ueno, D., Takahashi, S., Tanaka, H. et al. (2003). Global pollution monitoring of PCBs and organochlorine pesticides using skipjack tuna as a bioindicator. Archives of Environmental Contamination and Toxicology, 45, 378– 389. Unal, G. & Jepson, P. C. (1991). The toxicity of aphicide residues to beneficial invertebrates in cereal crops. Annals of Applied Biology, 118, 493– 502.


Underwood, A. J. (1994). On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecological Applications, 4, 3–15. van der Werf, H. M. G. (1996). Assessing the impact of pesticides on the environment. Agriculture, Ecosystems and the Environment, 60, 81–96. van Wijngaarden, R. P. A., Brock, T. C. M. & van den Brink, P. J. (2005). Threshold levels for effects of insecticides

in freshwater ecosystems: a review. Ecotoxicology, 14, 355–380. Williams, P., Whitfield, M., Biggs, J. et al. (2002). How realistic are outdoor microcosms? A comparison of the biota of microcosms and natural ponds. Environmental Toxicology and Chemistry, 21, 143–150.


Chapter 17

Assessing pesticide risks to humans: putting science into practice Brian Hughes, Larry G. Olsen and Fred Whitford IPM has long been a staple of decision making in agricultural settings. In recent years, the concept of IPM has been widely utilized in non-agricultural pest management programs. While the foundation of IPM has been to better manage pests, in many aspects, IPM has come to symbolize the management of pesticide risks. Thus, it seems that IPM is both a pest management and an integrated pesticide management program. IPM practitioners today use the IPM principles to predict and manage the risks that both pests and pesticides pose to people and the environment (see Chapter 37). Here we will address how the risks of pesticides are evaluated under USA federal statutes and how mitigation measures can reduce risk.

17.1 Regulatory framework Many of the older pesticides were metals (e.g. arsenic or copper) or compounds derived from plants (e.g. nicotine). Concerns were expressed that some of these substances could end up in processed food leading to food supply contamination. In 1906, the USA Congress passed the Food and Drugs Act to address the issue of adulterated products entering the food supply. This Act while general in language was quite powerful in that it became unlawful to manufacture a food product if it contained any unwanted or deleteri-

ous ingredients that could be injurious to human health. Congress would add to and clarify the Food and Drug Act over the years to update the law to meet growing concerns over food safety. A landmark event occurred in 1938 when the US Congress passed the Food, Drug, and Cosmetic Act which set the benchmark as it related to pesticides in food and setting general standards for reducing the risks of exogenous chemicals in and on the food supply. The basic foundation to the law was that federal agencies would have to set tolerances for pesticides used to produce food and fiber. The tolerance is the permissible residue level for pesticides in raw agricultural products or processed food. In 1954, the Miller Pesticide Amendment listed procedures for setting these safety limits for pesticide residues on raw agricultural commodities. Government, specifically the US (Food and Drug Administration) (FDA), was now controlling the levels and risk. Another federal law in the USA that tried to define risk was the Food Additives Amendment to the Food, Drug, and Cosmetic Act in 1958. What would become one of the most controversial aspects of the law was the Delaney clause. This clause prohibited food additives to be found in processed food if they were shown to induce cancer in humans or in laboratory animals. Pesticides were considered food additives to processed foods. There arose an inconsistency in these laws

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


which granted on one hand a tolerance for a pesticide on raw agricultural commodities based on a risk/benefit analysis versus a no tolerance for the pesticide that might cause cancer if found in processed food. The next major regulatory reform was a sweeping piece of legislation known as Food Quality Protection Act (FQPA) of 1996 (Food Quality Protection Act of 1996 § 1, 7 U.S.C. § 101). The law eliminated the inconsistency posed by the Delaney clause by excluding pesticides from the definition of food additives. It created a single standard for pesticides in both raw and processed foods. It greatly strengthened risk evaluation by considering combined exposures resulting from diet, drinking water and residential uses instead of evaluating each risk separately. With the FQPA, the exposures from each of these routes are added and compared to the toxicity data as the means of determining risk. The only exception is that occupational risks currently are still evaluated separately. Beyond setting tolerances for aggregate exposures, FQPA required the US Environmental Protection Agency (EPA) to consider cumulative effects of classes of pesticides or those having the same mechanism of toxicity. Thus, exposures over multiple active ingredients will be used in some cases to determine if those exposures impact human health. Another hallmark of FQPA was the incorporation of an additional safety factor for infants and children. The safety factor was intended to protect subpopulations that had special sensitivities as reported by the National Academy of Sciences Committee on Pesticide Residues in the Diets of Infants and Children in 1993 (National Academy of Sciences, 1993). FQPA requires that up to an additional ten-fold safety factor be incorporated into the risk assessment process to take into account pre- and postnatal toxicity shown in laboratory animals.

the environment. The risk assessment process has three overarching components: toxicity assessment/hazard identification, exposure assessment and risk characterization. The end result of risk assessment is the fourth step called risk management where EPA evaluates the risk(s), and determines whether additional steps need to be undertaken to reduce the risks to acceptable levels.

17.2.1 Toxicity assessment/hazard identification This assessment requires a variety of controlled experiments using laboratory animals which are used to determine the biological consequences of exposure to that pesticide. These tests are conducted by the product manufacturer or registrant and involve subjecting animals to a range of doses, through various routes and durations to determine the toxicological profile of the chemical. A key question to answer is at what level a pesticide produces no adverse effects.

17.2 Risk assessment

Acute toxicity Acute toxicity studies are used to mimic shortterm exposures to relatively high levels of the pesticide. These studies help to determine the adverse effects when the pesticide is taken by an oral, inhalation or dermal route. These are the initial studies that provide the underlying data that are used to develop the hazard assessment of a pesticide. These data help identify what problems arise out of an exposure duration of a day or less. They serve as one of the benchmarks that mimic occupational exposures to a pesticide. These studies provide information to address how occupational exposures resulting from early re-entry into fields, mixing/loading or application measure against levels that produce adverse effects such as skin irritation, blindness and death. They form the basis for later evaluations to determine what personal protective equipment will be required on the pesticide label. On the basis of these acute studies, pesticides are then classified by signal words as Danger (highly toxic), Warning (moderately toxic) and Caution (slightly toxic) (Table 17.1).

The EPA has an established risk assessment process by which pesticides are evaluated to determine their level of risks to human health and

Chronic toxicity Pesticide exposures occur over longer periods of time especially if one considers dietary, including




Table 17.1

Classification of pesticide toxicity


Signal word


Danger Warning Caution Caution

Oral LD50 (mg/kg) 5000

Dermal LD50 (mg/kg)

Inhalation LC50 (mg/l)

20 000


Source: Adapted from 40 Code of US Government Federal Regulations Part 156.62 (US Federal Government, 2003). drinking water, or residential contributions. Therefore, EPA requires the manufacturer to conduct a series of studies that measure the adverse impacts of subchronic (e.g. 90 days) studies and lifelong studies that mimic exposure over a lifetime (e.g. chronic studies). Subchronic studies involve exposing multiple animal species for up to 90 days through the same routes − oral, dermal and inhalation − in a similar manner as the acute studies. Chronic studies are used to determine the effects of a chemical to prolonged and repeated exposures covering much of the laboratory animal’s lifespan. Chronic studies provide the data that addresses the role of the pesticide to cause cancer, liver disorders, etc. Reproductive effects Reproductive studies are an important set of studies that allow one to address the potential impact on the developing fetus and the mother. These studies determine if the pesticide has any deleterious effects on the ability of adult animals to conceive and reproduce or the offspring to grow and develop. Mutagenicity The last battery of tests that help to describe the toxicological properties of a pesticide is one that measures its mutagenetic potential. This testing is usually conducted in vitro with the pesticide and/or its metabolites to determine if it induces mutations in bacterial or mammalian DNA. The data obtained in these studies are reviewed with the other studies that are designed to measure the pesticide’s oncogenicity and teratogenicity potential.

Lastly, a number of toxicokinetic studies define how the body handles the pesticide regarding absorption (oral, dermal or through inhalation), distribution (where does the pesticide go in the body), metabolism (how is the parent chemical altered) and excretion (how is the chemical eliminated). The preceding tests are those basic to the understanding of the toxicology of a pesticide. They are conducted according to published guidelines by the EPA. In some cases EPA will ask the manufacturer to conduct additional tests when the agency believes that additional toxicological data is needed to better describe at what levels adverse effects of a pesticide are observed. Laboratory studies that describe the toxicologic properties are used in lieu of using human subjects. Recently, the EPA outlined various protections for subjects in human research involved in pesticide testing. These protections included a ban on using sensitive subpopulations such as children or women of reproductive age and enhanced oversight of human studies conducted to support registration (US Federal Government, 2006). The hesitancy of using human subjects involve ethical considerations such as weighing the risk to the participants to the benefits they gain as a result of conducting the research as well as methodological considerations such as how much confidence to place on the low number of volunteers used in such studies. Another direct or indirect look at human health effects associated with pesticide exposure is through the use of epidemiological studies. Epidemiology is an observational science that looks at events within a population to see if there are strong or weak associations between


a population’s exposure to a pesticide and an adverse effect within that population. Two basic studies are used to determine whether a pesticide exposure can be linked to a health impact. A cohort study is conducted by selecting a group of people with similar exposures to a pesticide versus a control group of individuals who do not have the exposure. The point of demarcation is whether one is exposed or not. The epidemiologists will collect health information over time and into the future from the subjects to determine whether the incidence of disease in the exposed group is significantly different than the disease in the unexposed control group. The other classic epidemiology study is called a case-control study. In this method, an epidemiologist will start out with individuals with a disease and those without the disease. The scientist then goes back in time to gather detailed histories from each subject including diet, smoking and chemical exposures. In this study, attempts are made to link any exposures that might have resulted in higher disease rates.

17.2.2 Exposure assessment From the toxicology studies one can determine at what levels adverse effects occur in laboratory animals. However, these numbers in themselves have little meaning without knowing the full extent and magnitude of the exposure. Thus, an additional step of assessing exposure is needed to address the risk that a pesticide might pose. Exposure assessment asks the questions: how are humans exposed, how frequent is the exposure and what is the amount of exposure? While toxicology studies can be controlled, the answers to these questions asked in exposure studies are rather complex because they must include the variations associated with human behavior. In general, two assessments are conducted. A general exposure assessment is made to describe intake levels from pesticides through diet, drinking water and use around and in the home. The second is an occupational assessment where one evaluates agricultural workers that either contact pesticides in the field post-application or while performing tasks using pesticides.

General exposure Diet represents the primary source of exposure for most pesticides. The nature and extent of pesticide exposure through food and meat varies with the type of food, the pesticides used to produce those products, and the amount of food ingested. The basic estimate of dietary exposure is represented by the following equation: Pesticide intake in diet  = (Concentration of residue on each food × Amount of each food consumed) The food consumption data and pesticide residue information used in the equation is derived from multiple sources. The simplest evaluation is when EPA uses the legal tolerance levels that are set for the pesticide. Tolerances are the upper legal amount of a pesticide allowed on each crop (e.g. maize, spinach, tomato). These are called worstcase scenarios in that seldom are tolerance levels found on food products. It is an overestimation of the amount on food products. In other instances, data on pesticide residues that are actually on the food supply can be obtained from the US Department of Agriculture (USDA) Pesticide Data Program (PDP) (US Department of Agriculture, 2005). PDP is a national pesticide database program that has been in existence since 1991. Each year this program publishes a detailed report of their findings of how much pesticide is actually measured on food. Each year, new priorities are established that direct which foods are to be examined in greater detail. For instance, foods and juices consumed by infants and children have been given a high priority by the USDA. The data provides a wealth of residue (e.g. exposure) information for risk assessment calculations, such as ranges, means and lower limits of detection. Even the PDP can overestimate levels of pesticides consumed especially in those products that are cleaned and cooked. The Total Diet Study (TDS), popularly called the market basket study, is an ongoing FDA program that determines levels of various pesticides, contaminants and nutrients in foods after they are prepared for consumption. A unique aspect of the TDS is that foods are prepared as they would be consumed (table-ready) prior to analysis. It allows for the pesticide to be washed




off the vegetables and fruits as well as taking into account the destructive forces of heat in breaking down the pesticide. While it provides realistic estimates of how much pesticide is actually consumed, a single study can cost $US 1 million to conduct. The amount of pesticide residue on the food is but half of the equation. Obviously, the more one eats of a specific crop, the more potential there is for greater consumption of pesticides when it is found on that food. Food consumption data is critical in the calculation of the amount of pesticide residue consumed in the diet. The USDA is the primary agency assigned the responsibility for collecting food consumption data for the USA population. Since 1989, the consumption data has been collected annually through a program called the Continuing Survey of Food Intake by Individuals. Personal interviews with volunteers determine what the subject consumed within a 24-hour period. As expected, the American diet has drastically changed over the past decades. Within the last 20 years, increased emphasis is also given to the unique dietary habits of children since their proportionally greater intake of fruit and fruit juices may lead to increased intakes of pesticides used on these commodities. Since 2001, the USDA partnered with the National Health and Nutrition Examination Survey (NHANES) has worked to improve and conduct the collection, assessment and dissemination of food consumption and related data of Americans. This current survey is called the What We Eat In America (WWEIA) survey. The data collected is based on such descriptions as pizza. The pizza is then broken down into the various raw product components such as tomato, maize, green pepper, etc. The researchers can then break down each product into an amount consumed. The EPA considers drinking water to be an exposure pathway for certain pesticides and it is considered part of the aggregate assessment. The difficulty in assessing pesticides in drinking water is due to seasonal variations in the levels of contamination found in surface and groundwater. Various agencies monitor drinking water and have provided the data to the EPA. The most notable of these is the US Geological Survey’s National Water Quality Assessment Program. Generally speaking, the amount of pesticide exposure

is rather small as compared to the amount consumed in the diet. In other situations, the EPA will use various exposure models to predict the amount of estimated exposure of a pesticide in drinking water. Residential and non-occupational exposures may be among the most difficult estimates to predict; scientists need to examine, for example, exposure from homes, school and daycare centers. Use of pesticides in and around the home, on turf or on pets provides multiple points of exposure. For example, the general population also is exposed through how a homeowner manages pests in and around the home. The indoor environment provides opportunity for exposure from insect control like foggers and crack and crevice treatments. Also, flea and tick treatments are often applied to pets and transferred to humans. Sanitizers and disinfectants are also commonly used in the home. Indoor assessments are highly complicated due in part to the pesticides used, their application methods and human populations exposed. There are a number of factors that must be used when evaluating and calculating residential exposures:

(1) Residential building factors There are many types of residential construction: mobile homes, houses and apartments. Each varies with regard to the type of construction materials and the heating, air conditioning and ventilation requirements. (2) Demographic factors Each age group differs with respect to their residential exposures. Infants and toddlers have greater inhalation rates for their size. (3) Human activity patterns Infants and toddlers are more likely to crawl on floors near areas where a pesticide has been applied and concentrations are the greatest (Fenske et al., 1990). They also spend a greater amount of their lives indoors compared to working adults who spend significant amount of time away from home. The young and the elderly may be considered more highly exposed due to the lack of mobility. Adults may also handle and apply pesticides, particularly to lawns and gardens. Children on the other hand may play in these


areas incurring greater exposure to the skin and they may even eat treated foliage. The outdoor environment also has its unique exposures. Outdoors, weed and insect control products may be applied to lawns, gardens, trees and sidewalks to control pest problems around the house. Algaecides are used in swimming pools, paint and wood preservatives to protect decking, termiticides to protect structures and rodenticides to control mice and rats. Each use allows for exposures through use or human interactions in the environment. Turf applications are a major outside source of pesticide residue to the homeowner including the handling and application of pesticides. Foliar residues are a potential source of dermal pesticide exposure. Residues that can be transferred from foliage to humans are termed dislodgeable residues. The type and duration of human activity determine the residue concentrations transferred during normal human contact with treated surfaces. Dislodgeable residue studies are taken from foliage immediately after the applied residues have dried and for several days after to determine the decay or dissipation rate of the pesticide.

Occupational exposure The EPA is required under the Code of Federal Regulation, Title 40 (40 CFR 158.202) to establish minimum re-entry times before agricultural workers can re-enter agricultural fields after a pesticide has been applied. Beyond requiring toxicity data, the EPA requires studies involving dissipation of foliar pesticide and residue exposure to farm workers from post-application entry. These studies have been outlined by the EPA in their Occupational and Residential Exposure Guidelines. In 1994, the Agricultural Re-entry Task Force was organized to develop a generic agricultural re-entry exposure database intended to address agricultural post-application/re-entry data requirements. Passive whole-body dosimetry and inhalation exposure monitoring are the standard methodologies used to determine exposure levels. The EPA determined that exposure to pesticide handlers is more a function of the job and application technique rather than the specific product being applied.

Generic re-entry exposures were calculated using a set of crop groupings. Exposures are considered similar in these crop groupings if the plants are similar in structure, height and appearance. Thus, the plants were grouped according to characteristics they shared with each other (e.g. height, leaf structure, etc.) and agronomic activities involved (e.g. harvesting, thinning, pruning, scouting, etc.). Thus, maize would be a tall row crop which includes sorghum and sunflower as crops with similar physical characteristics. Exposures are also defined based on the types of activity involved in crop production. This information can be obtained from crop profiles. Crop profiles are descriptions of crop production and pest management practices compiled on a state basis for specific commodities and include worker activities that occur during the growing season. (Crop profiles can be found at: www. These profiles provide a tool that the EPA can use to identify low to high risk activities. Figure 17.1 illustrates the types of information available from crop profiles. Intuitively, exposure might be more if the activity required lots of contact when the plants are in the later growth stages (e.g. harvesting) as compared with activities when the plants are smaller and there is less plant surface area to come into contact. The end result is to predict potential exposures based on duration and activity of the workforce. For instance, workers involved in seed maize production can be divided into eight distinct activities and durations and ranked with their relative exposure potentials (Table 17.2). The important question is how much residue would dislodge itself from the crop onto the person walking through or working in a field. True exposure to the person is not how much is on the plant at the time of application, but how much of the residue remaining over time can be rubbed off the plant onto the person while performing various activities. Residue dissipation studies are conducted taking foliar samples and determining the dislodgeable foliar residue. Dislodgeable foliar residues (DFRs) are often described as residues present on the surface of a leaf that are available for transfer from the leaf onto skin or clothing. Foliar samples are taken prior to the day of application, directly after application (4 or




Table 17.2

Maize seed activities, duration and relative exposure potentials

Activity Planting Isolation Irrigation Pest scouting Rogueing Detasseling Certification of detasseling Phytosanitary inspection

Duration (weeks)

Exposure potential

5 6 14 12–16 6 6 6 8

Low Very low Very low Moderate to high Moderate Moderate to high Moderate to high Moderate to high

Source: L. G. Olsen (personal communication).

Fig. 17.1 Timeline indicating when events and activities occur in the production of maize seed (L. G. Olsen, personal communication).


Monitor breathing zone for inhalation of pesticide residues Passive Dosimetry: Amount of pesticide trapped on materials outside the body is quantified.

Active compound at target protein, tissue, etc.

Pesticide residue accumulates on body and is absorbed.

Biological Monitoring of Urine: Amount of pesticide or a metabolite is quantified in urine samples.

Absorbed pesticide is metabolized and distributed throughout the body

Fig. 17.2 Routes of exposure.

12 hours as appropriate) and at 1, 2, 4, 7, 10, 14, 21, 28 and 35 days after application. Foliar samples consist typically of a leaf punch that can accurately measure the recommended 400 cm2 of leaf surface to be processed and analyzed. The samples are taken randomly within each field and at varying heights in the crop. The processing of the plant material requires the leaf surface to be washed with an appropriate chemical to remove dislodgeable residues. The testing process requires the leaf tissue to be processed within 4 hours. The results are reported in µg of residue/cm2 of plant material. The data from the dislodgeable foliar residue analysis is used to calculate the transfer coefficient (TC). The TC is the amount of foliage that a worker comes into contact within 1 hour and is calculated by the following equation. Transfer coefficient (TC) = Dermal exposure (µg/hr)/DFR(µg/cm2 ) Transfer coefficients are dependent on the type of crop, application method and worker activity in the field. The application of this TC is to calculate the dose or dermal exposure of an individual. Dose (mg/kg per day) = (TC)(DFR)(AT)(AB)/(BW)

where TC = transfer coefficient in cm2 /h DFR = dislodgeable foliar residue in µg/cm2 AT = activity time in h AB = dermal absorption (fraction absorbed through the skin) BW = body weight in kg. The result is a dose in milligrams/kg body weight which can be compared to the appropriate RfD (see discussion under “Threshold effects” in Section 17.2.3). Dermal exposure from post-pesticide application is assessed by estimating two main points of body entry: inhalation and dermal contact. Estimates of dermal exposure often use the patch testing, whole body dosimetry, and hand rinse and face wipes. The external estimate of exposure is also known as passive dosimetry although some have used biological monitoring to estimate absorbed dose (Fig. 17.2). Patch testing involves placing 10–12 absorbent cloth patches on the outside of clothing on the chest, back, upper arm, forearm, thigh and lower leg. Each patch covers an approximate area of 100 cm2 . Residues are trapped and collected on the




cloth. At the end of a predetermined exposure period, the patches are analyzed for the presence of the pesticide. The resulting data collected for each patch is in µg of pesticide residue/cm2 of patch material. These results of how much pesticide was collected per cm2 are then extrapolated to determine how much pesticide would have been collected on the entire outside surface of the body. For example, a pesticide applicator is found to have 0.1 µg/cm2 from patches placed on the chest. This 0.1 µg/cm2 value would be multiplied by 3454 cm2 which is the average size of a male chest. Thus, it would be predicted that 345.4 µg of the pesticide would be expected to be found on the front chest. The EPA will also accept a similar test method known as whole body dosimetry. Whole-body dosimeters generally consist of an inner layer such as long underwear garments covered with long pants and a long-sleeved shirt of an absorbent material (Fig. 17.3). These clothes are worn throughout the exposure period while the person is involved in conducting field tasks such as planting, hoeing, weeding and harvesting. The inner dosimeter or undergarments are sectioned into upper and lower leg, upper and lower arm, and front and rear torso. Each piece of the dosimeter is analyzed separately and then are added together to obtain a total amount of residue. The outside layer also is analyzed for the amount of transferable residue. Often, these studies can show that the outer layer proves very difficult for many pesticides to move through onto the inner garments. The hands are often the part of the body which receives the majority of the pesticide exposure to the whole body. To estimate this amount of exposure on the skin surface, the subject is asked to wash their hands after they have performed their various jobs for a specified period of time. After the allotted time has expired, workers are asked to wash their hands with a solution known to remove the product from the surface of the skin. The wash solution is collected and analyzed. Gloves too can be used to capture this data in much the same way that whole-body dosimeters are used. Areas of the face and neck are assessed by wiping the area with a moist absorbent material and then submitting the material for analysis.

Fig. 17.3 Whole-body outer dosimeters.

Personal air samplers are used to estimate the amount of pesticide inhaled by workers. A portable battery powered monitoring pump is clipped to the belt and a tube is run up the back and clipped to the collar of each worker. Inserted into the tube near the collar is an absorbent filter that traps airborne pesticides. The filter is then analyzed to determine the amount of pesticide per liter of airflow.


For mixer/loaders and applicators, industry and the regulatory agencies cooperated in establishing the Pesticide Handlers Exposure Database (PHED) in the early 1990s to gather data for occupational risk assessments for these activities. Companies contributed their exposure data for the common benefit of all registrants. PHED was then used by the regulatory agencies and registrants to satisfy the data requirements. PHED satisfied the need for occupational exposure data during the 1990s. However, PHED is incomplete since many exposure scenarios are not covered and today there are new handling systems and formulations that are not covered in PHED. Similar to the Agricultural Re-entry Task Force, industry and the EPA formed the Agricultural Handlers Exposure Task Force to gather a generic database to determine exposures to pesticide mixer, loader and applicators in support of pesticide registration.

17.2.3 Risk characterization The integration of data from the toxicity studies compared to the estimates of exposure is a detailed process called risk characterization. Risk assessors conducting these evaluations often describe risks in numerical terms. This risk characterization is necessary to determine if actions are required to increase the margin of safety to the general population or agricultural workers from pesticide use. Therefore, the greater the extent to which the exposure is below an adverse effect level, then that product is less risky. Concerns are raised when the estimates of exposure are above levels known to have produced a recognizable symptom. Making conclusions around a single line – exposures below known toxicity levels are “safe” − are in themselves risky statements. There are many reasons why one has concerns about simple models such as these. There are variables that are hard to define. For instance, might there be differences in response between laboratory animals and people? Might individuals have different sensitivities compared to others resulting in different responses like between children and adults? Thus, risk characterization errs on the side of safety. The EPA scientists take into account the uncertainties in the data and the unanswered

questions by reducing the toxicological line that separates one level that creates an adverse effect from the one that doesn’t. For instance, the toxicological data indicates that at 100 mg/kg of body weight the pesticide does not produce an observable effect. Instead, the EPA might divide the level by a factor of 100 to derive a level of limited concern at 1 mg/kg. The exposure level is then compared to the lower level which has the built-in 100 × safety factor. The threshold for safety in this example is increased by 100-fold. The risk characterization adequately takes into account the quality of the data in terms of its variability and uncertainty and how each is accounted for in the assessment.

Threshold effects A threshold effect is one where certain levels produce an adverse effect, while lower levels do not. For threshold effects only, the EPA determines a Reference Dose (RfD) or acceptable daily intakes in which the agency can reasonably anticipate the level found in food is unlikely to harm the consumer. RfDs are based on oral intakes for various periods of exposure (acute, subchronic and chronic) in different species of laboratory animals. For example, one may establish a No Observable Adverse Effect Level (NOAEL) is 10 mg/kg per day based on a 2-year (chronic) dietary study in rats. To establish an RfD, the EPA will divide the NOAEL by various safety factors. The safety factors include 10× based on the uncertainty for interspecies extrapolation from animals to humans or another 10× based on the variability in the human population. Additional safety or modifying factors may be incorporated into the RfD including 3× to 10× based on database sufficiency or as directed by FQPA where children and infants need additional protection. Dietary and drinking water evaluations depend on the RfDs as comparison values. For instance, the NOAEL from the laboratory animals is 10 mg/kg per day. With the uncertainties inherent in the risk characterization, the assessor could divide by a safety factor of 300 (10× for interspecies, 10× for human variability and 3× for children and infants). Thus, whether a pesticide is characterized as risky depends on




whether the exposure that is estimated goes above 0.003 mg/kg per day. RfD = NOAEL/Safety factors

In the assessment of occupational risk, a Margin of Exposure (MOE) is developed by dividing the NOAEL by the exposure estimate. The difference between the two values is obviously a percentage. The EPA uses the guideline of a MOE of 100 or greater as a margin of safety to assure that worker exposures will not result in any adverse health effects. The acceptable margin of safety is developed similar to that of the RfD safety factor approach: 10× to account for the uncertainty in extrapolating from effects seen in animal studies to humans and 10× to account for the variable responses within the human population to pesticides. Non-threshold (carcinogenic) effect It is believed that cancers and some chronic diseases can result from accumulation of adverse effects resulting from low-level exposures over a period of time and may not be evident until some time after the exposures have occurred. In other words, these diseases do not operate in terms of a threshold creating the problem. The EPA assesses the effects of pesticides differently to calculate the risk of exposure and the development of cancer. The medical community and the EPA make the assumption that any exposure can be accompanied by an increase risk of cancer. Therefore, a risk of 1 in 1 million indicates that if everyone in a population of 1 million were similarly exposed to a chemical, there may be one additional cancer case in that population. The level of acceptable risk to the general population is usually 1 in 1 million but may vary as a policy decision between various state and federal agencies and the affected population. Dose–response curves from animal studies are used to estimate carcinogenic risk. The lower end of the dose response curve is extrapolated back to zero using various mathematical models. Since this mathematical construct is an estimate, a dose–response curve that is at the upper 95th confidence level is used for risk assessment purposes. The use of upper 95th percentile indicates that the true dose–response curve will be represented 95%

of the time. Thus, at a risk of 1 in 1 million there is a 95% certainty that at most only one additional cancer case in a population will result and may be zero. The slope of the dose-response curves at the 95% confidence level represents cancer potency. This is called the q∗ with units of reciprocal mg/kg body weight per day or (mg/kg body weight)−1 . Risk is a unitless value calculated by finding the dose absorbed multiplied by the q∗ . Aggregate risk In the past, risk characteristics were often made separately for dietary, residential, water intake and occupation. FQPA requires that the EPA calculate the total dose from pesticides through all routes of exposure. In other words, the EPA must take all of the exposures from food, water and residential use for a specific pesticide to determine how that level compares to some toxicological endpoint. For dietary risk, the total dose is delivered through residues on a wide variety of fruits and vegetables all through the oral route. Therefore, the evaluation of dietary risk and toxicology endpoints derived from an animal study involving oral administration of the pesticide is straightforward. The process for determining aggregate risk is more complicated because other activities and routes of exposure are summed. For instance, a child’s risk is dietary and also residential which involves exposure to treated surfaces in the home and turf outside the home. So the aggregate to a child involves oral exposure not only to foods but also from hand-to-mouth behaviors, inhalation and dermal exposures in and around the home. The aggregate must also be specific to the duration of the exposure. If an exposure is acute, then only data obtained from acute exposure studies should be used in the assessment. Second, the studies must match with respect to their route of exposure; if the exposure is oral, dermal or inhalation then the route of exposure in the animal study should match. Should the data not be available for the specific route of exposure then the data can be used to extrapolate an absorbed dose. For instance, where exposure is on the skin and no reliable dermal study can be found, then


it can be assumed that a certain percentage of the dose is absorbed through the skin. In this way the risks are calculated, then summed as follows: Dietary (i.e. eating an apple): Ia = C a × R a

where Ca = consumption rate (in g/kg body weight per day) Ra = residue (µg of pesticide/grams of food consumed) Ia = intake (µg of pesticide/kg of body weight per day).

17.2.4 Risk management The goal of the risk assessment process is to provide a mechanism to manage pesticide risks. During the risk management process the EPA determines the risk and mitigation measures needed to reduce food, water, worker and environmental risks to acceptable levels. The worst-case scenario for pesticides that do not meet that standard would be not to register that active ingredient, or the cancellation of the right of the manufacturer to sell their product in the market. Often this is not a good solution for either the producer or the consumer particularly when there are few alternate options to control pests safely and economically. One option is for the registrant to generate more realistic numbers for input into the risk assessment process. For food residues that exceed the RfD, the EPA can take action to reduce the allowable uses on different commodities. With each use removed or denied, the theoretical exposure to that pesticide is reduced. Furthermore, the EPA can also limit the number of applications of a pesticide or limit the amount per application to ensure such levels are met. Pre-harvest intervals can also be lengthened to allow time for residue dissipation in the field. For pest control in the home, many insecticides are now in the form of gels and baits. Application techniques have also changed so that pesticides applied indoors via wide area applications are now crack and crevice treatments which reduce the amount of pesticide and the potential for exposures. In many states, laws prevent the pesticide

from being used in or around institutional settings like schools and daycare centers when buildings or grounds are being occupied. Pesticide use can also be made less risky by repackaging or reformulating products or by adding dyes or bitterants. Workers may be exposed upon re-entering fields after pesticide application. The EPA can manage the risks to these workers by limiting the number of applications or reducing the rate of application or increasing the Restricted Entry Interval (REI). Consider the case in Table 17.3 where the various MOEs for azinphos methyl have been calculated for almonds and apples. The MOEs have been calculated using the labeled rate and the particular TC for each activity listed. Based on the former REIs the MOEs are less than 100. To manage the risks the REIs have been lengthened to where the calculated MOEs meet the target level of 100. For those mixing/loading or applying pesticides, personal protective equipment (PPE) is required by label. EPA has instituted worker protection standards to protect workers by requiring PPE, central notification, availability of sanitary facilities and posting requirements restricting entry after pesticide application. Risk management decisions may also take into account the availability of other control methods. For example, the availability of other chemistries where the toxicity is lower can result in a greater MOE. Newer chemistries can include the use of neonicotinoids or insect growth regulators. Plant incorporated protectants such as the Bacillus thuringiensis Bt maize can lower the number of pesticide applications and the problems with timing of application.

17.3 Using pesticide risk information to make IPM decisions The EPA and the manufacturers identify the level of risks posed by pesticides not just as an exercise for product registration for the marketplace, but to provide safety information. The identification of the product’s toxicological characteristics is just as much a part of the decision making




Table 17.3

Margins of exposure (MOEs) for post-application agricultural re-entry activities

Crop (maximum label rate per application)

Transfer coefficients for each activity


Almonds (2.0 lbs ai/A)

2 day REL: MOE = 3 for irrigating and scouting 14 day REL: N/A (no hand thinning) 28 day PHI: MOE = 3 for poling mummy nuts and pruning; REL where the MOE reaches 100: 71 days for irrigating, scouting & hand weeding, 104 days for poling & pruning

TC = 400 cm2 /h for irrigating, scouting and hand weeding TC = 2500 cm2 /h for poling and pruning

Apples, crab apples (1.5 lbs ai/A)

2 day REL: MOE = 23 for propping; MOE = 2 for irrigation and scouting; MOE = 1 for pruning, tying and training 14 day REL: MOE = 1 for hand thinning 14/21 day PHI: MOE = 2 for hand harvesting. REL where the MOE reaches 100: 32 days for propping; 79 days for irrigating, scouting and weeding; 102 days for hand harvesting, hand thinning, pruning, tying and training

TC = 100 cm2 /h for propping TC = 1000 cm2 /h for irrigating, scouting and weeding TC = 3000 cm2 /h for hand harvesting and thinning, pruning, tying and training

Source: Data from Environmental Protectioin Agency (2001). process for growers as is using insect thresholds, disease forecast models, product efficacy and scouting programs. Registration of a new active ingredient involves years of research and millions of dollars to bring the risk side into balance with the benefits side. This is important because the risk assessment process provides pesticide applicators reduced risk alternatives to reach IPM goals. IPM professionals see themselves as environmental problem-solving consultants who use pesticides more judiciously than in the past with the goal of minimizing the occurrence of pests using the lowest risk combination of tools appropriate to the individual situation. To reach these goals, IPM professionals are challenged to develop new approaches to manage insects, rodents, plant diseases and weeds. No matter how diligent the IPM practitioner is in the use of alternative management strategies, they will find that pesticides (conventional or organic) are an important tool in their IPM program.

An IPM practitioner can use the information developed for the EPA registration process to identify products that pose less risk to people and environment. The following criteria are useful in the selection of a pesticide product for use in an IPM program. (1) What is the signal word on a pesticide label? The relative acute toxicity of a pesticide product is reflected on the label by one of three signal words: Danger − most toxic, Warning − moderately toxic, Caution − least toxic. The signal word can also reflect the product’s nonlethal effects such as skin and eye irritation. Products in the Caution category are preferred for IPM programs when a pesticide is required. (2) Is the product classified for restricted use or general use? Compounds that are classified by EPA as restricted-use products (RUP) are analogous to prescription drugs that must be administered by a trained person such as a doctor or nurse. On the other hand, general-use


pesticides (everything not a RUP) are like overthe-counter medicines that any person can purchase and use. A pesticide product is classified as an RUP due to concerns about the potential for harm if the product is not used properly. RUPs are only sold to and used by persons with certification issued by the state agency which is responsible for pesticide regulation. Although general-use pesticides are generally less risky than the RUPs, both can cause harm if not used according to label directions. RUPs are not used in IPM programs if alternatives exist. (3) Is the product a reduced-risk pesticide? The EPA has created a category of pesticides that pose fewer risks when compared to other products currently in the marketplace for the control of similar pests. To obtain EPA’s reduced-risk classification manufacturers must submit substantive data. Claims must be supported by evidence of reduced toxicity to humans or to other non-target organisms and/or improved environmental fate and transport. In return, EPA expedites the registration process for these products. (4) Does the product have acute and/or chronic effects? Manufacturers produce a Material Safety Data Sheet (MSDS) for every pesticide active ingredient. The MSDS provides a great deal of information about the hazards of a product including acute and chronic effects, carcinogenicity and reproductive/developmental defects. It is important to remember that most MSDS are produced for the concentrated product in the jug or bag. Therefore the MSDS serves as a hazard communication document mainly for occupational use and the risks when the product is diluted are typically lower than for the concentrated form in the bag or jug. Selecting reduced risk products in concert with limiting exposure through sound application methods is essential to sound IPM practices. The product label in essence is the ultimate risk management document used to minimize any risks associated with pesticide use and still allow users to reap its benefits. It is more than a list of directions on how to control specific pests − product costs, pests controlled, application rates and crops it can be used on. It

also carries a tremendous amount of information based on the human health, wildlife protection and environmental testing process put in place by the EPA. For instance, it provides valuable information on preharvest application requirements that must be followed to prevent unacceptable residues on raw and processed foods. Another example is the reentry interval that prevents individuals from returning to fields after a pesticide application to protect themselves from unacceptable exposures.

17.4 Conclusions Many IPM professionals are providing what customers want, service that maximizes pest control while minimizing pesticide exposure to the environment and to applicators, farmers and clients. IPM represents a major step forward for growers, licensed pesticide applicators and the firms for which they work. It is the natural progression for growers and companies focused upon risk management as well as pest management. As governmental agencies and customers struggle with the positive and negative aspects of pesticides in every day life, IPM provides a “middle path” for a holistic and integrated approach to manage pest problems coupled with responsible pesticide use. Clients appreciate and often are willing to pay a premium for a program managed by a professional trained and skilled in the tenets of IPM. Pest management service providers understand that the best IPM programs include more narrowly focused selection and application of reduced risk pesticides, and better trained, more observant and more skilled advisors.

References Fenske, R. A., Black, K. G., Elkner, K. P. et al. (1990). Potential exposure and health risks of infants following indoor residential pesticide applications. American Journal of Public Health, 80, 689–693. National Academy of Sciences (1993). Pesticides in the Diets of Infants and Children. Washington, DC: National Academies Press.




US Department of Agriculture (2005). Pesticide Data Program: Annual Summary Calendar Year 2005. Washington, DC: US Department of Agriculture, Agricultural Marketing Service. Environmental Protection Agency (2001). Interim Reregistration Eligibility Decision for Azinphos-Methyl Case No. 0235, October 2001. Washington, DC: US Environmental Protection Agency, Office of Prevention, Pesticides and Toxic Substances. US Federal Government (2003). 156.62 Toxicity category. Code of Federal Regulation, Title 40 CFR Chapter 1 (7–1-03 edition). Available at http://a257.g. pdf. US Federal Government (2006). Final rule. Protections for subjects in human research. February 6, 2006. Federal Register: US Environmental Protection Agency (EPA), 71, 6137–6176. Whitford, F. (2002). The Complete Book of Pesticide Management: Science, Regulation, Stewardship, and Communication. New York: Wiley Interscience. Whitford, F., Pike, D., Burroughs, F. et al. (2006). The Pesticide Marketplace: Discovering and Developing New Products, Purdue Pesticide Programs No. PPP-71. West Lafayette, IN: Purdue University.

Chapter 18

Advances in breeding for host plant resistance C. Michael Smith Production of crop plants with heritable arthropod resistance traits has been recognized for more than 100 years as a sound approach to crop protection (Painter, 1951; Smith, 2005). Hundreds of arthropod-resistant crops are grown globally and represent the results of long-standing cooperative efforts of entomologists and plant breeders. These crops significantly improve world food production, increase producer profits and contribute to reduced insecticide use and residues in food crops (Smith, 2004). It is essential to determine the inheritance of arthropod resistance genes. Plant breeders do so by observing progeny segregating from crosses between resistant and susceptible parents to determine the mode of inheritance and action of the resistance gene or genes. Breeding methods such as mass selection, pure line selection, recurrent selection, backcross breeding and pedigree breeding are often used to incorporate arthropod resistance genes into cultivars of such crops as maize, rapeseed, rice, wheat, potato, cotton and alfalfa (Smith, 2005). The focus of this chapter is on how the inheritance of resistance has been determined for the development of these crops and how new methods have been adapted in twentieth- and twenty-first-century plant breeding to select for arthropod resistance genes.

18.1 Inheritance of resistance Khush & Brar (1991) and Gatehouse et al. (1994) have prepared extensive reviews on the inheritance of arthropod resistance in food and fiber crops. The following subsections summarize the number of genes involved in resistance and the mode of inheritance of these genes to pests of fruit, forage, oilseed and vegetable crops, as well as the major cereal crops of maize, rice, sorghum and wheat. Specific reference citations in this section are listed in the supplemental bibliography.

18.1.1 Fruit, forage, oilseed and vegetable crops Resistance in apple, lettuce, peach and raspberry to several species of aphids is controlled by the action of one or two genes inherited as dominant traits. Crosses between cultivated potato and different Solanum species have also shown that glandular trichome-based resistance to the green peach aphid (Myzus persicae) is a partially dominant trait controlled by one or two dominant genes. Similarly, resistance to the potato tuber moth (Phthorimaea operculella) is controlled by a small number of major genes.

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



Resistance in mung bean (Vigna radiata) to the Azuki bean weevil (Callosobruchus chinensis) and the cowpea weevil (Callosobruchus maculatus) as well as cowpea (Vigna unguliculata) resistance to cowpea aphid (Aphis craccivora) are all also controlled by single genes inherited as dominant traits. The phytomelanin layer in Helianthus species resistant to sunflower moth (Homoeosoma electellum) is also inherited as a dominant trait controlled by a single gene. Resistance in common bean (Phaseolus vulgaris) to Mexican bean weevil (Zabrotes subfasciatus) is controlled by a toxic seed protein (arcelin) and is also inherited as a dominant trait. However, resistance to a bruchiid (Acanthoscelides obtectus) also derived from a Phaseolus accession, is controlled by two genes, inherited as recessive traits. Resistance in tomato to the spider mite Tetranychus evansi is also controlled by the action of one gene that segregates as a recessive trait. Resistance in soybean to several species of defoliating Lepidoptera is linked to quantitative trait loci (QTLs). Similarly, resistance in common bean and lima bean to a leaf hopper (Empoasca kraemeri) and the inheritance of hooked trichomes, an Empoasca resistance mechanism, are linked to QTLs. In the forage legumes alfalfa (Medicago sativa) and sweetclover (Melilotus infesta) single dominant genes control resistance to pea aphid (Acyrthosiphon pisum) and sweetclover aphid (Therioaphis riehmi). In contrast, alfalfa resistance to spotted alfalfa aphid (Therioaphis maculata) is controlled by QTLs.

18.1.2 Maize Multiple genes linked to QTLs control resistance in maize to larval defoliation and stalk damage inflicted by several species of Lepidoptera. These include corn earworm (Helicoverpa zea), European corn borer (Ostrinia nubilalis), stem borer (Sesamia nonagrioides), spotted stem borer (Chilo partellus), southwestern corn borer (Diatraea grandiosella) and sugarcane borer (D. saccharalis). Depending on the pest, gene action involves epistatic as well as additive-dominance effects. Different genes condition resistance to both first and second generations of the European corn borer, but some genes condition resistance to both generations.

18.1.3 Sorghum Sorghum resistance to greenbug (Schizaphis graminum) was originally found to be controlled by a single gene inherited as a partially dominant trait. QTL analyses have more recently been used to document resistance to different greenbug biotypes as well as resistance to the sorghum midge (Stenodiplosis sorghicola). In some genotypes of sorghum resistant to the sorghum shootfly (Atherigona soccata) resistance is controlled by QTLs and also inherited as a partially dominant trait. In other genotypes expressing resistance based on leaf trichomes, resistance is expressed as a recessive trait conditioned by a single gene. Two genes inherited as recessive traits control resistance in sorghum to head bug (Eurystylus oldi). Resistance to another pentatomid, Calocoris angustatus, is inherited as a partially dominant trait controlled by both additive and nonadditive gene action.

18.1.4 Rice Many genes have been identified in rice or related wild relatives for resistance to a complex of pests including brown planthopper (Nilaparvata lugens), green rice leafhopper (Nephotettix cincticeps), green leafhopper (Nephotettix virescens), Asian rice gall midge (Orseolia oryzae), whitebacked planthopper (Sogatella furcifera) and zigzag leafhopper (Recilia dorsalis) (Table 18.1). Four genes control resistance to green rice leafhopper, and QTLs for each are located on four different chromosomes. Eight genes control expression of resistance in rice to green leafhopper. Six are inherited as dominant traits and two genes are inherited as recessive traits and linked to QTLs. Four genes inherited as dominant traits and one gene inherited as a recessive trait control whitebacked planthopper resistance in rice, while three different dominant genes condition resistance to zigzag leafhopper. In the most extensively studied system, 13 genes control resistance to brown planthopper. Of these, six genes are inherited as dominant traits and seven genes are inherited as recessive traits. QTLs associated with antixenosis and tolerance resistance to brown planthopper have been identified. Rice gall midge resistance is also an


Table 18.1

Genes in rice and related wild relatives controlling resistance to leafhoppers, planthoppers and the

rice gall midge


Resistance genesa

Green rice leafhopper Nephotettix cincticeps

Grh1, 2, 3, 4

Green leafhopper Nephotettix virescens

Glh1, 2, 3, 5, 6, 7 glh4, 8

Brown planthopper Nilaparvata lugens

Bph1, 3, 6, 9, 10, 13 bph2, 4, 5, 7, 8, 11, 12

Rice gall midge Orseolia oryzae

Gm1, 2, 4, 5, 6, 7, 8, 9 gm3

Zigzag leafhopper Recilia dorsalis Whitebacked planthopper Sogatella furcifera

Z1h1, 2, 3

a b

Wbph1, 2, 3, 5 wbph4

Referencesb Kobayashi et al., 1980; Saka et al., 1997; Fukuta et al., 1998; Yazawa et al., 1998; Tamura et al., 1999; Wang et al., 2003 Athwal & Pathak, 1971; Siwi & Khush, 1977; Rezaul Kamin & Pathak, 1982; Pathak & Khan, 1994; Wang et al., 2004 Athwal & Pathak, 1971; Lakashminarayana & Khush, 1977; Ikeda & Kaneda, 1981; Kabir & Khush, 1988; Ishii et al., 1994; Kawaguchi et al., 2001; Renganayaki et al., 2002 Satyanarayanaiah & Reddi, 1972; Sastry & Praska Rao, 1973; Chaudhary et al., 1986; Sahu & Sahu, 1989; Tomar and Prasad, 1992; Srivastava et al., 1993; Yang et al., 1997; Kumar & Sahu, 1998; Kumar et al., 2000a, b; Katiyar et al., 2001; Shrivastava et al., 2003 Angeles et al., 1986 Sidhu et al., 1979; Angeles et al., 1981; Hernandez & Khush, 1981; Wu & Khush, 1985

Uppercase – inherited as a dominant trait; lowercase – inherited as a recessive trait. References provided in the supplemental online bibliography (Smith, 2007).

extensively studied gene-for-gene system in arthropod resistance. Eight different genes inherited as dominant trait genes and one monogenic recessive gene have been documented as controlling gall midge resistance in India and China. Indepth discussions of rice–arthropod gene-for-gene interactions involving Asian rice gall midge and brown planthopper are provided in Chapter 12 of Smith (2005).

18.1.5 Wheat Genes from barley, rye and wheat wild relatives have been transferred into bread wheat to provide resistance to numerous arthropod pests (see review by Berzonsky et al., 2003). The most extensively studied pest is Hessian fly (Mayetiola destructor), which began infesting wheat in the USA in the first decade of the twentieth century. More than

25 rye or wheat genes control Hessian fly resistance and all but one (h4) are inherited as dominant or partially dominant traits (Table 18.2). The deployment of these resistance genes in response to Hessian fly biotypes is also discussed in Chapter 12 of Smith (2005). Seven genes expressing resistance to five biotypes of the greenbug have been characterized. Gb2, Gb3, Gb5, Gb6, Gbx, Gby and Gbz are all inherited as single dominant traits. Similarly, ten genes in wheat and two genes in barley confer resistance to Russian wheat aphid (Diuraphis noxia). All but one, the recessive gene dn3, are inherited as dominant traits. Resistance to the wheat curl mite (Aceria tosichella), a vector of wheat streak mosaic virus, is also controlled by four different genes from rye or various wheat wild relatives. All are inherited as dominant traits.




Table 18.2

Genes in wheat expressing resistance to arthropod pests


Resistance genesa

Wheat curl mite Aceria tosichilla

Cm1, 2, 3, 4

Russian wheat aphid Diuraphis noxia

Dn1, 2, 4, 5, 6, 7, 8, 9, x, dn3, Rdn1, Rdn2

Greenbug Schizaphis graminum

Gb2, 3, 5, 6, x, y, z,

Hessian fly Mayetiola destructor

H1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 30, h4

a b

Referencesb Thomas & Conner, 1986; Schlegel & Kynast, 1987; Whelan & Hart, 1988; Whelan & Thomas, 1989; Chen et al., 1996; Cox et al., 1999; Malik et al., 2003 du Toit, 1987, 1988, 1989; Nkongolo et al., 1989, 1991a, b; Harvey & Martin, 1990; Marais & du Toit, 1993; Marais et al., 1994, 1998; Mornhinweg et al., 1995, 2002; Saidi & Quick, 1996; Ma et al., 1998; Zhang et al., 1998; Liu, 2001; Liu et al., 2002 Livers & Harvey, 1969; Sebesta & Wood, 1978; Harvey et al., 1980; Joppa et al., 1980; Hollenhorst & Joppa, 1983; Tyler et a1., 1987; Porter et al., 1994; Weng & Lazar, 2002; Boyko et al., 2004; Zhu et al., 2004 Stebbins et al., 1982, 1983; Maas et al., 1987; Ratcliffe & Hatchett, 1997; Martin-Sanchez et al., 2003

Uppercase – inherited as a dominant trait; lowercase – inherited as a recessive trait. References provided in the supplemental online bibliography (Smith, 2007).

18.2 Resistance gene clusters Several of the maize QTLs mentioned above that play major roles in resistance to the European corn borer, the southwestern corn borer and the sugarcane borer, occur on maize chromosomes 2, 5, 7 and 9. The major QTLs for production of maysin and apimaysin, flavonoid allelochemicals controlling resistance to feeding by the corn earworm, also occur on chromosomes 5 and 9. The relationships between the QTLs for the different types of resistance on chromosomes 5 and 9 have not been investigated. Genes for resistance to arthropod pests occur in clusters in maize, rice and wheat. In rice, four brown planthopper resistance genes map to a block on chromosome 12 and two genes map to a second cluster on chromosome 3. In wheat, five genes controlling Russian wheat aphid resistance are located in a cluster on the

short arm of wheat chromosome 7D, and five genes for resistance to greenbug are clustered on the distal portion of the long arm of chromosome 7D (Fig. 18.1). More recently, QTL on both the long and short arm of wheat chromosome 7D have been shown to play a role in resistance to Russian wheat aphid and greenbug (Castro et al., 2004).

18.3 Resistance gene mapping High-density genetic maps of nearly all major agricultural crops such as barley, maize, potato, rye, sorghum, soybean, tomato and wheat have been developed (see review by Smith, 2005), and molecular markers in many of these crops are linked to genes expressing resistance to several major arthropod pests (see review by Yencho et al., 2000). Using these resources, genetic mapping techniques allow comparison of a plant phenotype and a plant genotype. Plant genotypes are


Fig. 18.1 Aphid resistance gene loci in sorghum and wheat on Triticeae homoeologous chromosome group 7. Left side of chromosome: Dn – single dominant genes from wheat for Diuraphis noxia resistance, Gb – single dominant genes from wheat for Schizaphis graminum resistance, Ssg – major restriction fragment length polymorphism (RFLP) loci from sorghum controlling S. graminum resistance. Right side of chromosome: RFLP loci spanned on far right by disease defense response loci or resistance gene analog loci (as indicated). Positions of loci are not ordered. For complete description and discussion see Smith (2005).




determined after the amplification of plant DNA with multiple molecular markers of known chromosome location, and estimates of the genetic linkage between the resistance gene and specific markers are then determined. Many molecular markers linked to single resistance genes inherited as dominant traits have been identified. Multiple markers linked to groups of QTLs controlling resistance have also been identified. The markerassisted selection of plants based on genotype, before phenotypic resistance is determined, is becoming more common. Linkage between resistance genes and molecular markers varies greatly. They may be completely linked, where no crossing-over occurs between the gene and the marker during meiosis, or incompletely linked, with crossing-over between the two. Genes and markers may have no linkage, because they are located on separate chromosomes or far from one another on the same chromosome. Estimates of the recombination between a resistance gene and a linked marker are measured as the recombination frequency, which is measured in segregating plant populations by pairing the phenotype and genotype of each progeny and analyzing these data with computer software such as Mapmaker or Mapmaker/QTL (Lincoln et al., 1993). To estimate recombination frequency, DNA is collected from tissues of resistant and susceptible parent plants and 100–200 F2 plants or plants from 100–200 F2 -derived F3 families of known resistance or susceptibility. Molecular markers from multiple chromosome locations are screened to identify those producing polymorphisms between the DNA of the parents, and if parent DNA banding pattern polymorphisms exist, the loci of a marker is said to be informative of the resistance gene location (Fig. 18.2). Two DNA samples, one from several highly resistant and one from several highly susceptible plants, are then amplified with the informative marker. If parent polymorphisms exist between the bulked resistant and susceptible DNA samples, the marker is putatively linked to the gene, DNA of all F2 plants or F3 families is amplified, and phenotype and genotype data are subjected to Mapmaker analysis.

Fig. 18.2 Banding patterns of DNA from leaves of wheat plants lines amplified with a simple sequence repeat (SSR) marker of known wheat chromosome location. R, resistant phenotype parent; S, susceptible phenotype parent; L, 100 base pair DNA reference ladder. Arrow, putative resistance-specific DNA band. Amplification products were electrophoresed in a 3% agarose gel stained with ethidium bromide (from Flinn, 2000).

18.4 Molecular markers The use of molecular markers has advantages compared to morphological markers. Some molecular markers behave in a co-dominant manner to detect heterozygotes in segregating progeny when morphological markers detect dominant or recessive traits. In addition, the allelic variation detected by molecular markers is considerably greater than that detected by morphological markers, and molecular markers are unaffected by environmental affects. Molecular markers used to determine arthropod resistance gene location include restriction fragment length polymorphism (RFLP) markers, random amplified polymorphic (RAPD) markers, amplified fragment length polymorphism (AFLP) markers and simple sequence repeats (SSRs) or microsatellite markers. RFLP markers detect differences between genotypes when restriction enzymes cut genomic DNA to yield variable-sized DNA fragments that are then separated by electrophoresis. Digested DNA is transferred to a nylon membrane (Southern blotting) and the membrane is probed with a


radioactive labeled, single-stranded DNA probe of known chromosome location. Membrane-bound DNA is denatured by heat and the probe sequences bind to complementary sites in the restriction digest. Unbound probe is removed by washing and the dried membrane is exposed to x-ray film and photographically developed as an autoradiogram. Binding between probe and membrane-bound DNA provides information about the location of a resistant gene in the form of different (polymorphic) autoradiogram DNA banding patterns between two genotypes. When restriction sites in the vicinity of a gene are compared between genotypes, one genotype may have the site, while the other does not. If differences exist, they are referred to as polymorphisms between the two genotypes. As indicated above, RFLP markers detect heterozygotes and have been used to map arthropod resistance gene loci in numerous crops (Yencho et al., 2000). The disadvantages of RFLP linkage analysis include the 7 to 10 day time period required to complete an analysis and the use of radioactive isotopes. Polymerase chain reaction (PCR) primers from known chromosome locations are reacted with template DNA, amplified in a thermal cycler, and the amplification products are electrophoresed to identify primers yielding polymorphic banding patterns between resistant and susceptible plants. Compared to RFLP hybridization, PCR amplification is many times faster and does not require the use of radioactive materials. Several types of PCR primers have been used to identify plant resistance genes. The use of AFLP markers involves digesting DNA with different restriction enzymes and annealing restriction enzyme adaptors to the restriction products. Restriction digests are preselected by PCR amplification with general restriction enzymes attached to unique oligonucleotide primers. Preselected PCR products are then selectively amplified using specific oligonucleotide primers, amplified fragments are separated by electrophoresis and gel images are converted to autoradiograms. AFLP markers have been used to successfully map arthropod resistance genes in apple, rice and wheat.

Simple sequence repeat (SSR) or microsatellite PCR primers are two to five base dinucleotide repeats widely distributed in eukaryotic DNA. Microsatellite primers generate high levels of polymorphism, detect patterns of co-dominant inheritance and have been used to map several arthropod resistance genes in rice and wheat (Yencho et al., 2000). Comparison of RFLP, AFLP and SSR markers in cultivated and wild soybean revealed a high correlation between the three marker types (Powell et al., 1996). Linkage between QTLs and marker loci is based on the relation between the phenotypic expression of several minor resistance genes and molecular markers at multiple loci. QTL analyses determine which loci explain the greatest amount of phenotypic variation for a biochemical or biophysical character controlling resistance. QTLs are linked to arthropod resistance genes in barley, maize, potato, rice, sorghum, soybean, tomato and wheat (see reviews by Yencho et al., 2000 and Smith, 2005).

18.5 Molecular marker-assisted selection Cost and labor savings have been documented for the use of a molecular marker-assisted selection system based on SSR markers linked to genes for nematode resistance in soybean and barley (Mudge et al., 1997; Kretshmer et al., 1997) and for disease resistance in rice (Hittalmani et al., 2000; Toenniessen et al., 2003). In arthropod resistance research, the use of maker-assisted selection continues to increase as a strategy for single-gene resistance inherited as a dominant trait. An online bibliography provides additional references on marker-assisted selection (Smith, 2007). For QTLs in general, these markers are yet to be used to their fullest extent in marker-assisted selection (see reviews of Babu et al., 2004 and Food and Agriculture Organization, 2003). The same situation exists for arthropod resistance studies. In a study to determine transfer of soybean QTLs for Lepidoptera resistance into cultivars, Narvel et al. (2001) found very few resistant genotypes




with multiple QTLs from different soybean linkage groups, and suggested marker-assisted selection to introgress QTLs for resistance into elite germplasm. Results of research with maize QTLs linked to Lepidoptera resistance further illustrate the limited use of QTLs in marker-assisted selection. Comparisons of marker-assisted selection and phenotypic selection of maize resistance to southwestern cornborer and sugarcane borer by Groh et al. (1998), Bohn et al. (2001) and Willcox et al. (2002) demonstrated that the efficiency of both methods was similar, and suggested that phenotypic selection is more favorable, due to reduced costs. Environmental variation also affects the use of QTLs for Diatrea resistance marker-assisted selection (Groh et al., 1998). Thus, as noted in these above studies, the effective use of marker-assisted selection for Diatrea resistance will depend on major markerassisted selection cost reductions with this technique, the development of more QTLs that explain more variance for resistance and the expression of QTLs over a broad range of environments.

The practice of releasing a resistant cultivar containing a single major gene, planting it until it becomes ineffective and making additional sequential releases of other major genes is quite common in rice and wheat insect pest resistance breeding programs. Pyramiding (incorporation of more than one resistance gene) of two or more major genes in one cultivar, although time consuming, increases longevity of resistance genes and has been used successfully to protect rice cultivars with brown planthopper resistance. In the case of greenbug resistance in wheat, pyramiding provides no additional protection over that provided by single resistance genes released sequentially. Finally, the development and deployment of multiline cultivars composed of different combinations of major and minor resistance genes has been used for rice resistance to brown plant hopper and green rice leafhopper, wheat resistance to Hessian fly and sorghum resistance to sorghum midge.

18.7 Conclusions 18.6 Deploying insect resistance genes As reviewed above, resistance inherited as either single genes (monogenic) or QTLs (polygenic) has been deployed in many different insect pest management systems. This has occurred in spite of the fact that monogenic resistance often results in the development of virulent biotypes of insects that are unaffected by the original plant resistance gene. Comparatively, polygenic resistance is often considered more stable than monogenic resistance and is not readily overcome by resistance-breaking arthropod biotypes. The polygenic nature of maize resistance to lepidopteran larvae is a good example of this relationship. Nevertheless, the majority of the insect-resistant cultivars, including transgenic plants possessing insecticidal proteins, contain single-gene resistance. This is because plants containing single traits are easier to score and the population size necessary to study the inheritance of resistance is smaller than for the evaluation of polygenic resistance.

Insect-resistant crops will continue to play a very important role in world sustainable agricultural systems. It is very likely that the benefits of insectresistant crops will become more acute as world climate change increases and food needs and food availability become more uncertain, especially in the underdeveloped and developing countries of the semi-tropics. Although conventional and transgenic resistance breeding efforts have made major strides to improve maize, rice and wheat during the past century, the important food crops of the semi-arid tropics crops (sorghum, millet, pigeon pea and chickpea) remain in need of identification and deployment of increased amounts of insect resistance. Refinement and increased use of marker-assisted selection techniques should be encouraged in order to accelerate the rate and accuracy of breeding of all crop plants for insect resistance. The continued evolution of virulent biotypes dictates the need for identification of new sources of resistance and heightens need for marker-assisted selection systems to identify and track these genes.


Genomic technologies have opened completely new avenues of research in plant resistance to insects over the last decade. Genomic microarrays of several crop plants (barley, maize, rice, tomato, wheat) are beginning to provide critical information about the identity of resistance genes, their chromosome location and the gene products mediating the function of resistance genes. The sequencing of the genomes of rice and the model plant Arabidopsis are providing insights into plant insect resistance gene structure, function and location. The soon to be completed sequencing of maize and the model legume Medicago truncatulata will provide additional insights as well. As more plant genomes are sequenced, existing and new information about resistance gene synteny can be used to make foresighted decisions about the design and breeding of insect resistant crop plants. Knowledge about the diversity of resistance genes will permit breeding of crop cultivars with resistance genes of diverse sequence and function that will help delay the development of resistance-breaking insect biotypes. Sequence information afforded by resistance gene analogs in many crop plants will also allow plant resistance researchers to use this in silico resource to determine more in-depth information about location of candidate resistance genes and the biochemical and biophysical gene products mediating their function. The ultimate goal of resistance gene expression studies, genomic studies, and marker-assisted selection systems should be to identify plant genes that can be cloned and used to transform crop plants for durable insect resistance.

References Angeles, E. R., Khush, G. S. & Heinrichs, E. A. (1981). New genes for resistance to whitebacked planthopper in rice. Crop Science, 21, 47–50. Angeles, E. R., Khush, G. S. & Heinrichs, E. A. (1986). Inheritance of resistance to planthoppers and leafhoppers in rice. In Rice Genetics, Proceedings of the International Rice Genetics Symposium, May 27–31, 1985, pp. 537–549. Manila, Philippines: Island Publishing Company. Athwal, D. S. & Pathak, M. D. (1971). Genetics of resistance to rice insects. In Rice Breeding, pp. 375–386. Los Ba˜ nos, Philippines: International Rice Research Institute.

Babu, R., Nair, S. K., Prasanna, B. M. & Gupta, H. S. (2004). Integrating marker-assisted selection in crop breeding: prospects and challenges. Current Science, 87, 607– 619. Berzonsky, W. A., Ding, H., Haley, S. D. et al. (2003). Breeding wheat for resistance to insects. Plant Breeding Review, 22, 221–296. Bohn, M., Groh, S., Khairallah, M. M. et al. (2001). Re-evaluation of the prospects of marker-assisted selection for improving insect resistance against Diatraea spp. in tropical maize by cross validation and independent validation. Theoretical and Applied Genetics, 103, 1059–1067. Boyko, E. V., Starkey, S. R. & Smith, C. M. (2004). Molecular genetic mapping of Gby, a new greenbug resistance gene in bread wheat. Theoretical and Applied Genetics, 109, 1230–1236. Castro, A. M., Vasicek, A., Ellerbrook, C. et al. (2004). Mapping quantitative trait loci in wheat for resistance against greenbug and Russian wheat aphid. Plant Breeding, 123, 229–332. Chaudhary, B. P., Srivastava, P. S., Shrivastava, M. N. & Khush, G. S. (1986). Inheritance of resistance to gall midge in some cultivars of rice. In Rice Genetics, pp. 523–528. Los Ba˜ nos, Philippines: International Rice Research Institute. Chen, Q., Conner, R. L. & Laroche, A. (1996). Molecular characterization of Haynaldia villosa chromatin in wheat lines carrying resistance to wheat curl mite colonization. Theoretical and Applied Genetics, 93, 679– 684. Cox, T. S., Bockus, W. W., Gill, B. S. et al. (1999). Registration of KS96WGRC40 hard red winter wheat germplasm resistant to wheat curl mite, stagonospora leaf blotch, and septoria leaf blotch. Crop Science, 39, 597. du Toit, F. (1987). Resistance in wheat (Triticum aestivum) to Diuraphis noxia (Hemiptera: Aphididae). Cereal Research Communications, 15, 175–179. du Toit, F. (1988). Another source of Russian wheat aphid (Diuraphis noxia) resistance in Triticum aestivum. Cereal Research Communications, 16, 105–106. du Toit, F. (1989). Inheritance of resistance in two Triticum aestivum lines to Russian wheat aphid (Homoptera: Aphididae). Journal of Economic Entomology, 82, 1251–12153. Food and Agriculture Organization (2003). Molecular Marker Assisted Selection as a Potential Tool for Genetic Improvement of Crops, Forest Trees, Livestock and Fish in Developing Countries. FAO Electronic Forum on Biotechnology in Food and Agriculture. Available at




Flinn, M. (2000). A molecular marker linked to tolerance in Aegilops tauschii Accession 1675 to greenbug (Homoptera: Aphididae). Manahattan, KS: M.S. Thesis, Kansas State University. Fukuta, Y., Tamura, K., Hirae, M. & Oya, S. (1998). Genetic analysis of resistance to green rice leafhopper (Nephotettix cincticeps Uhler) in rice parental line, Norin-PL6, using RFLP markers. Breeding Science, 48, 243–249. Gatehouse, A. M. R., Boulter, D. & Hilder, V. A. (1994). Potential of plant-derived genes in the genetic manipulation of crops for insect resistance. Plant Genetic Manipulation for Crop Protection, 7, 155–181. Groh, S., Gonzalez-deLeon, D., Khairallah M. M. et al. (1998). QTL mapping in tropical maize. III. Genomic regions for resistance to Diatraea spp. and associated traits in two RIL populations. Crop Science, 38, 1062– 1072. Harvey, T.L. & Martin, T. J. (1990). Resistance to Russian wheat aphid, Diuraphis noxia, in wheat (Triticum aestivum). Cereal Research Communications, 18, 127– 129. Harvey, T. L., Martin, T. J. & Livers, R. W. (1980). Resistance to biotype C greenbug in synthetic hexaploid wheats derived from Triticum tauschii. Journal of Economic Entomology, 73, 387–389. Hernandez, J. E. & Khush, G. S. (1981). Genetics of resistance to whitebacked planthopper in some rice (Oryza sativa L.) varieties. Oryza, 18, 44–50. Hittalmani, S., Parco, A., Mew, T. W., Zeigler, R. S. & Huang, N. (2000). Fine mapping and DNA markerassisted pyramiding of the three major genes for blast resistance in rice. Theoretical and Applied Genetics, 100, 1121–1128. Hollenhorst, M. M. & Joppa, L. R. (1983). Chromosomal location of genes for resistance to greenbug in ‘Largo’ and ‘Amigo’ wheats. Crop Science, 23, 91–93. Ikeda, R. & Kaneda, C. (1981). Genetic analysis of resistance to brown planthopper, Nilaparvata lugens St˚ al, in rice. Japan Journal of Breeding, 31, 279–285. Ishii, T., Brar, D. S., Multani, D. S. & Khush, G. S. (1994). Molecular tagging of genes for brown planthopper resistance and earliness introgressed from Oryza australiensis into cultivated rice, O. sativa. Genome, 37, 217– 221. Joppa, L. R., Timian, R. G. & Williams, N. D. (1980). Inheritance of resistance to greenbug toxicity in an amphiploid of Triticum turgidum/T. tauschi. Crop Science, 20, 343–344. Kabir, M. A. & Khush, G. S. (1988). Genetic analysis of resistance to brown planthopper in rice (Oryza sativa L.). Plant Breeding, 100, 54–598.

Katiyar, S. K., Tan, Y., Huang, B. et al. (2001). Molecular mapping of gene Gm-6(t) which confers resistance against four biotypes of Asian rice gall midge in China. Theoretical and Applied Genetics, 103, 953–961. Kawaguchi, M., Murata, K., Ishii, T. et al. (2001). Assignment of a brown planthopper (Nilaparvata lugens St˚ al) resistance gene bph4 to the rice chromosome 6. Breeding Science, 51, 13–18. Khush, G. S. & Brar, D. S. (1991). Genetics of resistance to insects in crop plants. Advances in Agronomy, 45, 223– 274. Kobayashi, A., Kaneda, C., Ikeda, R. & Ikehashi, H. (1980). Inheritance of resistance to green rice leafhopper, Nephotettix cincticeps, in rice. Japan Journal of Plant Breeding, 30 (Suppl. 1), 56–57. Kretshmer, J. M., Chalmers, K. J., Manning, S. et al. (1997). RFLP mapping of the Ha2 cereal cyst nematode resistance gene in barley. Theoretical and Applied Genetics, 94, 1060–1064. Kumar, A. & Sahu, R. K. (1998). Genetic analysis for gall midge resistance: a reconsideration. Rice Genetics Newsletter, 15, 142–143. Kumar, A., Shrivasta, M. N. & Shukla, B. C. (2000a). Genetic analysis of gall midge (Orseolia oryzae Wood Mason) biotype 1 resistance in the rice cultivar RP 2333–156-8. Oryza, 37, 79–80. Kumar, A., Bhandarkar, S., Pophlay, D. J. & Shrivasta, M. N. (2000b). A new gene for gall midge resistance in rice accession Jhitpiti. Rice Genetics Newsletter, 17, 83–84. Lakashminarayana, A. & Khush, G. S. (1977). New genes for resistance to the brown planthopper in rice. Crop Science, 17, 96–l00. Lincoln, S. E., Daly, M. J. & Lander, E. S. (1993). MAPMAKER/EXP Version 3.0., A Tutorial and Reference Manual, 3rd edn. Cambridge, MA: a Whitehead Institute for Biomedical Research. Liu, X. (2001). Molecular mapping of wheat genes expressing resistance to the Russian wheat aphid, Diuraphis noxia (Mordvilko) (Homoptera: Aphididae). Manhattan, KS: Ph.D. dissertation, Kansas State University. Liu, X. M., Smith, C. M. & Gill, B. S. (2002). Mapping of microsatellite markers linked to the Dn4 and Dn6 genes expressing Russian wheat aphid resistance in wheat. Theoretical and Applied Genetics, 104, 1042–1048. Livers, R. W. & Harvey, T. L. (1969). Greenbug resistance in rye. Journal of Economic Entomology, 62, 1368–1370. Ma, Z.-Q., Saidi, A., Quick, J. S. & Lapitan, N. L. V. (1998). Genetic mapping of Russian wheat aphid resistance genes Dn2 and Dn4 in wheat. Genome, 41, 303– 306. Maas, F. B. III, Patterson, F. L., Foster, J. E. & Hatchett, J. H. (1987). Expression and inheritance of resistance


of ‘Marquillo’ wheat to Hessian fly biotype D. Crop Science, 27, 49–52. Malik, R., Smith, C. M., Harvey, T. L. & Brown-Guedira, G. L. (2003). Genetic mapping of wheat curl mite resistance genes Cmc3 and Cmc4 in common wheat. Crop Science, 43, 644–650. Marais, G. F. & du Toit, F. A. (1993). A monosomic analysis of Russian wheat aphid resistance in the common wheat PI294994. Plant Breeding, 111, 246–248. Marais, G. F., Horn, M. & du Toit, F. A. (1994). Intergeneric transfer (rye to wheat) of a gene(s) for Russian wheat aphid resistance. Plant Breeding, 113, 265–271. Marais, G. F., Wessels, W. G. & Horn, M. (1998). Association of a stem rust resistance gene (Sr45) and two Russian wheat aphid resistance genes (Dn5 and Dn7) with mapped structural loci in common wheat. South African Journal of Plant and Soil, 15, 67–71. Martin-Sanchez, J. A., Gomez-Colmenarejo, M., Del Moral, J. et al. (2003). A new Hessian fly resistance gene (H30) transferred from the wild grass Aegilops triuncialis to hexaploid wheat. Theoretical and Applied Genetics, 106, 1248–55. Mornhinweg, D. W., Porter, D. R. & Webster, J. A. (1995). Inheritance of Russian wheat aphid resistance in spring barley. Crop Science, 35, 1368–1371. Mornhinweg, D. W., Porter, D. R. & Webster, J. A. (2002). Inheritance of Russian wheat aphid resistance in spring barley germplasm line STARS-9577B. Crop Science, 42, 1891–1893. Mudge, J., Cregan, P. B., Kenworthy, J. P. et al. (1997). Two microsatellite markers that flank the major soybean cyst nematode resistance locus. Crop Science, 37, 1611– 1615. Narvel, J. A., Walker, D. R., Rector, B. G. et al. (2001). A retrospective DNA marker assessment of the development of insect resistant soybean. Crop Science, 41, 1931–1939. Nkongolo, K. K., Quick, J. S., Meyers, W. L. & Peairs, F. B. (1989). Russian wheat aphid resistance of wheat, rye, and triticale in greenhouse tests. Cereal Research Communications, 17, 227–232. Nkongolo, K. K., Quick, J. S., Limin, A. E. & Fowler, D. B. (1991a). Sources and inheritance of resistance to Russian wheat aphid in Triticum species amphiploids and Triticum tauschii. Canadian Journal of Plant Science, 71, 703–708. Nkongolo, K. K., Quick, J. S., Peairs, F. B. & Meyer W. L. (1991b). Inheritance of resistance of PI 373129 wheat to the Russian wheat aphid. Crop Science, 31, 905– 906. Painter, R. H. (1951). Insect Resistance in Crop Plants. Lawrence, KS: University of Kansas Press.

Pathak, M. D. & Khan, Z. R. (1994). Insect Pests of Rice. Los Ba˜ nos, Philippines: International Rice Research Institute. Porter, D. R., Webster, J. A. & Friebe, B. (1994). Inheritance of greenbug biotype G resistance in wheat. Crop Science, 34, 625–628. Powell, W., Morgante, M., Andre, C. et al. (1996). The comparison of RFLP, RAPD, AFLP and SSR (microsatellite) markers for germplasm analysis. Journal of Molecular Breeding, 2, 225–238. Ratcliffe, R. H. & J. H. Hatchett (1997). Biology and genetics of the Hessian fly and resistance in wheat. In New Developments in Entomology, ed. K. Bondari, pp. 47– 56. Trivandrum, India: Research Signpost, Scientific Information Guild. Renganayaki, K., Fritz, A. K., Sadasivam, S. et al. (2002). Mapping and progress toward map-based cloning of brown planthopper biotype-4 resistance gene introgressed Oryza officinalis into cultivated rice, O. sativa. Crop Science, 42, 2112–2117. Rezaul Kamin, A. N. M. & Pathak, M. D. (1982). New genes for resistance to green leafhopper, Nephotettix virescens (Distant) in rice, Oryza sativa L. Crop Protection, 1, 483– 490. Sahu, V. N. & Sahu, R. K. (1989). Inheritance and linkage relationships of gall midge resistance with purple leaf, apiculus and scent in rice. Oryza, 26, 79–83. Saidi, A. & Quick, J. S. (1996). Inheritance and allelic relationships among Russian wheat aphid resistance genes in winter wheat. Crop Science, 36, 256–258. Saka, N., Toyama, T., Tuji, T., Nakamae, H. & Izawa, T. (1997). Fine mapping of green ricehopper resistant gene Grh-3 (t) and screening of Grh-3 (t) among green ricehopper resistant and green leafhopper resistant cultivars in rice. Breeding Science, 47 (Suppl. 1), 55 (in Japanese). Sastry, M. V. S. & Prakasa Rao, P. S. (1973). Inheritance of resistance to rice gall midge Pachydiplosis oryzae Wood Mason. Current Science, 42, 652–653. Satyanarayanaiah, K. & Reddi, M. V. (1972). Inheritance of resistance to insect gall midge (Pachydiplosis oryzae, Wood Mason) in rice. Andhra Agriculture Journal, 19, 1–8. Schlegel, R. & Kynast, R. (1987). Confirmation of 1A/1R wheat–rye chromosome translocation in the wheat variety ‘Amigo’. Plant Breeding, 98, 57–60. Sebesta, E. E. & Wood, Jr., E. A. (1978). Transfer of greenbug resistance from rye to wheat with x-rays. Agronomy Abstracts, 1978, 61–62. Shrivastava, M. N., Kumar, A., Bhandarkar, S., Shukla, B. C. & Agrawal, K. C. (2003). A new gene for resistance in rice to Asian rice gall midge (Orseolia oryzae Wood




Mason) biotype 1 population at Raipur, India. Euphytica, 130,143–145. Sidhu, G. S., Khush, G. S. & Medrano, F. G. (1979). A dominant gene in rice for resistance to whitebacked planthopper and its relationship to other plant characteristics. Euphytica, 28, 227–232. Siwi, B. H. & Khush, G. S. (1977). New genes for resistance to the green leafhopper in rice. Crop Science, 17, 17– 20. Smith, C. M. (2004). Plant resistance against pests: issues and strategies. In Integrated Pest Management: Potential, Constraints and Challenges, eds. O. Koul, G. S. Dhaliwal & G. W. Cuperus, pp. 147–167. Wallingford, UK: CABI Publishing. Smith, C. M. (2005). Plant Resistance to Arthropods: Molecular and Conventional Approaches. Dordrecht, Netherlands: Springer-Verlag. Smith, C. M. (2007). Advances in Breeding for Host Plant Resistance: Supplemental Bibliography. Available at http://ipmworld/ Srivastava, M. N., Kumar, A., Shrivastava, S. K. & Sahu, R. K. (1993). A new gene for resistance to rice gall midge in rice variety Abhaya. Rice Genetics Newsletter, 10, 79– 80. Stebbins, N. B., Patterson, F. L. & Gallun, R. L. (1983). Inheritance of resistance of PI94587 wheat to biotypes B and D of Hessian fly. Crop Science, 23, 251–253. Stebbins, N. B., Patterson, F. L. & Gallun, R. L. (1982). Interrelationships among wheat genes H3, H6, H9, and Hl0 for Hessian fly resistance. Crop Science, 22, 1029–1032. Tamura, K., Fukuta, Y., Hirae, M. et al. (1999). Mapping of the Grh1 locus for green rice leafhopper resistance in rice using RFLP markers. Breeding Science, 49, 11– 14. Thomas, J. B. & Conner, R. L. (1986). Resistance to colonization by the wheat curl mite in Aegilops squarrosa and its inheritance after transfer to common wheat. Crop Science, 26, 527–530. Toenniessen, G. H., O’Toole, J. C. & DeVries, J. (2003). Advances in plant biotechnology and its adoption in developing countries. Current Opinion in Plant Biology, 8, 191–198. Tomar, J. B. & Prasad, S. C. (1992). Genetic analysis of resistance to gall midge (Orseolia oryzae Wood Mason) in rice. Plant Breeding, 109, 159–167. Tyler, J. M., Webster, J. A. & Merkle, O. G. (1987). Designations for genes in wheat germplasm conferring greenbug resistance. Crop Science, 27, 526–527.

Wang, C., Yasui, H., Yoshimura, A., Zhai, H. & Wan, J. (2003). Green rice leafhopper resistance gene transferred through backcrossing and CAPs marker assisted selection. Chinese Agricultural Science, 2, 13–18. Wang, C., Yasui, H., Yoshimura, A., Zhai, H. & Wan, J. (2004). Inheritance and QTL mapping of antibiosis to green leafhopper in rice. Crop Science, 44, 389–393. Weng, Y. & Lazar, M. D. (2002). Amplified fragment length polymorphism- and simple sequence repeatbased molecular tagging and mapping of greenbug resistance gene Gb3 in wheat. Plant Breeding, 121, 218– 223. Whelan, E. D. P. & Hart, G. E. (1988). A spontaneous translocation that transfers wheat curl mite resistance from decaploid Agropyron elongatum to common wheat. Genome, 30, 289–292. Whelan, E. D. P. & Thomas, J. B. (1989). Chromosomal location in common wheat of a gene (Cmc1) from Aegilops squarrosa that conditions resistance to colonization by the wheat curl mite. Genome, 32, 1033– 1036. Willcox, M. C., Khairallah, M. M., Bergvinson, D. et al. (2002). Selection for resistance to southwestern corn borer using marker-assisted selection and conventional backcrossing. Crop Science, 42, 1516–1528. Wu, C. F. & Khush, G. S. (1985). A new dominant gene for resistance to whitebacked planthopper in rice. Crop Science, 25, 505–509. Yang, D., Parco, A., Nandi, S. et al. (1997). Construction of a bacterial artificial chromosome (BAC) library and identification of overlapping BAC clones with chromosome 4-specific RFLP markers in rice. Theoretical and Applied Genetics, 95, 1147–1154. Yazawa, S., Yasui, H., Yoshimura, A. & Iwata, N. (1998). RFLP mapping of genes for resistance to green rice leafhopper (Nephotettix cincticeps Uhler) in rice cultivar DV85 using near isogenic lines. Science Bulletin of the Faculty of Agriculture at Kyushu University, 52, 169–175. Yencho, G. C., Cohen, M. B. & Byrne, P. F. (2000). Applications of tagging and mapping insect resistance loci in plants. Annual Review of Entomology, 45, 393–422. Zhang Y., Quick, J. S. & Liu, S. (1998). Genetic variation in PI 294994 wheat for resistance to Russian wheat aphid. Crop Science, 38, 527–530. Zhu, L. C., Smith, C. M., Fritz, A., Boyko, E. V. & Flinn, M. B. (2004). Genetic analysis and molecular mapping of a wheat gene conferring tolerance to the greenbug (Schizaphis graminum Rondani). Theoretical and Applied Genetics, 109, 289–293.

Chapter 19

Resistance management to transgenic insecticidal plants Anthony M. Shelton and Jian-Zhou Zhao Among the biological concerns expressed about the use of transgenic insecticidal plants are the potential for their genes to spread to wild and cultivated crops, have deleterious effects on nontarget organisms, and for insects to evolve resistance to the toxins(s) expressed in the plant (Shelton et al., 2002). These same concerns apply to many other forms of pest management, but transgenic plants have come under more scrutiny than most other strategies. For a broad review of the risks and benefits of insecticidal transgenic plants, the reader is referred to Sanvido et al. (2007). The focus of this chapter is solely on the risks of insects developing resistance to transgenic insecticidal plants and how such risks can be reduced. There is a long history of literature on resistance to conventional insecticides and to traditionally bred plants that are resistant to insects, and they provide a solid foundation for understanding the evolution and management of insect resistance to transgenic insecticidal plants. A commonly used definition of resistance is that it is a genetic change in response to selection by toxicants that may impair control in the field (Sawicki, 1987). Such changes could result from physiological or behavioral adaptation, although many more examples are from the former. Resistance to traditional synthetic pesticides has become one of the major driving

forces altering the development of IPM programs worldwide (Shelton & Roush, 2000). There are over 500 species of arthropods that have developed strains resistant to one or more of the five principal classes of insecticides (Georghiou & Lagunes-Tejeda, 1991), and that list continues to grow ( There are now well-documented cases of insects having developed resistance not only to synthetic insecticides, but also to pathogens including bacteria, fungi, viruses and nematodes (Shelton & Roush, 2000). Resistance has recently been found to spinosad and indoxacarb in the diamondback moth (Plutella xylostella) after less than 3 years of use (Zhao et al., 2006). This insect has a long history of rapidly becoming resistant to most insecticides used intensively against it, and is a harbinger of potential problems with other insects. This is especially relevant because it is the only insect that has developed high levels of resistance to proteins from the bacterium Bacillus thuringiensis (Bt), in the field, although this was to foliar sprays of Bt. Resistance is not one-dimensional. Not only do organisms have various physiological, biochemical and behavioral methods of developing resistance, but resistance has spatial and temporal components (Croft, 1990). Insects in one field may be susceptible to an insecticide while insects in

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



a nearby field may have developed high levels of resistance to that same insecticide (Shelton et al., 2006). Likewise, insects that were once resistant to an insecticide may regain susceptibility if that material is not used for some length of time, especially if fitness costs (e.g. lower fecundity) to the insect are associated with resistance (Croft, 1990). It is the variability in the spatial and temporal aspects of resistance to a particular insecticide (or class of insecticide) that can be exploited and allow the insecticide to be part of an overall insecticide resistance management (IRM) program. Reducing the selection pressure to a particular insecticide, and hence allowing it to be part of an IRM program, will delay the evolution of resistance and this should be the goal of any IRM strategy. To achieve this end will require an understanding of the genetics of resistance, monitoring the frequency of resistance alleles, and finding other tactics that will reduce the use of a particular insecticide while still providing adequate management of the insect population. Croft (1990) provides a list of examples of IRM programs for arthropods that have been “more successful” or “less successful.” Although more recent examples could be added to each category, the important message is that resistance can be managed and that the more successful programs have done so by: increased understanding of resistance mechanisms and enhanced resistance monitoring, increased cooperation of growers and government agencies and integration of IRM into an overall IPM program.

19.1 Insect-resistant plants Host plant resistance should be a foundation of IPM. There are over 100 insect-resistant crop cultivars grown in the USA and probably twice that many worldwide (Smith, 1989, ch. 18). The majority of these cultivars are field crops and the main target pests are sucking insects (aphids and leafhoppers). The benefits of genetically incorporated insect control can be substantial and include increased security to the grower, decreased use of insecticides and the potential to enhance biological control through conservation of natural enemies. However, like any technology directed

toward insect management, there is the risk that insects will evolve methods of overcoming the plant’s defenses. An important historical example illustrates this well: Hessian fly (Mayetiola destructor), a significant pest of wheat in the USA (see Gould, 1986). At least 16 wheat genes that confer resistance to this insect have been identified and a number of resistant wheat cultivars have been planted. However, their effectiveness has been diminished by Hessian fly biotypes that have developed resistance to the once-resistant plants. In many respects this is similar to the situation with traditional insecticides noted above. Considerable efforts have been made to breed resistance to Lepidoptera in many crops, but there have been few commercial successes outside of increasing the concentration of the cyclic hydroxamic acid DIMBOA in some cereal crops, most notably maize (Smith, 1989). DIMBOA affects the development of some insects and thus is classified as a form of antibiosis using Painter’s (1951) classification. DIMBOA has activity against some Lepidoptera as well as some sucking pests and can reduce the level of damage by the European corn borer (Ostrinia nubilalis) during the early growth stages of the maize. However, for protection of the ear, other strategies (most notably insecticide sprays) are used. No high level of host plant resistance to Lepidoptera or Coleoptera, orders that contain the most destructive insect pests of crops worldwide, has been developed and commercialized through conventional breeding methods. With the advent of biotechnology, breeders were no longer limited to genes from plants that could be used to produce insect-resistant plants, even to Lepidoptera and Coleoptera. The bacterium B. thuringiensis (Bt) offered some unique opportunities since different strains of most Bts contain varying combinations of insecticidal crystal proteins (ICPs), and different ICPs are toxic to different groups of insects (Tabashnik, 1994). Insecticidal products containing Bt subspecies were first commercialized in France in the late 1930s and dozens of Bt products have been sprayed on crops since. Bt genes that express specific ICPs were first introduced into tobacco plants for control of Lepidoptera in 1987 but more effective plants that used synthetic genes modeled on those


from Bt, but designed (plant codon optimized) to be more compatible with plant expression, were introduced a few years later (see review by Shelton et al., 2002). Bt plants displayed the first high level of host plant resistance to major lepidopteran pests and to some coleopteran pests. Of the $US 8100 million spent annually on all insecticides worldwide, it has been estimated that nearly $US 2700 million could be substituted with Bt biotechnology products for Lepidoptera (Krattiger, 1997). In 2007, 42.1 million ha of plants (maize and cotton) that express Bt proteins were utilized (James, 2007). Thus, Bt, which was once a fairly minor insecticide (99.9 97.3

0 2.3

0 0.4



No. of fields

1983 N.C. S.C.

6 600 21 240

800 2 000

1984 N.C. S.C.

9 200 32 000

1985 N.C. S.C. 1986 N.C. S.C.


Fig. 21.2 The number of hectares treated each year with codling moth commercial mating disruption products in Washington State, USA from 1990 to 2002 (Brunner et al., 2001; years 2001 and 2002 provided by Brunner, personal communication). In 1995 the CAMP program (Codling Moth Areawide Management Project) began with subsidized mating disruption applications provided to growers. The three CAMP sites in Washington State comprised a constant 760 hectares during the five years of the program (1995–1999).

an 80% or greater reduction of the use of broad-spectrum conventional insecticides by the end of the five-year program. A subsidy was provided to all participating growers of $US 125/ha for the first three years of the project to help defray the cost ($US 275/ha) of the mating disruption treatments (Brunner et al., 2001). For the final two years, growers had to pay the full cost of the treatments themselves. Standard highdose codling moth monitoring traps were used to assess trap-capture reductions and damage assess-

ments were made for all blocks within the CAMP project and compared with fruit from non-CAMPparticipating grower blocks. The number of hectares of apples under mating disruption in Washington State increased steadily during this project as word of successful codling moth population suppression spread (Fig. 21.2). Growers statewide continued to use this technique for years after the subsidies had disappeared, indicating that they were satisfied with the population suppression and their economic




Table 21.2

Pink bollworm damage to cotton bolls during successive years of the Parker Valley Mating Disruption Project, Arizona, USA

Larvae per 100 bolls

9 July 16 July 23 July 30 July 6 Aug 13 Aug 27 Aug 3 Sept 10 Sept 17 Sept Total bolls sampled/yr






– – 3.6 7.7 17.9 25.9 36.4 34.5 21.6 28.4

0.3 0.6 1.4 2.7 1.5 5.9 20.6 10.9 10.4 33.3

– 0.6 0.03 0.09 0.03 0.03 1.6 1.9 3.7 6.6

0.19 0.09 0.95 0.61 0 2.45 1.19 1.6 1.78 –

0 – – 0 0 0 0 0 0 0

23 847

31 630

21 675

25 603

22 852

balance sheets. Participants in the CAMP program achieved a 75% reduction in insecticide applications while reducing damage to unprecedented levels (Brunner et al., 2001). Damage at harvest fell from 0.8% the year before the program started to 0.55% during year 1, 0.2% during year 2 and to between 0.01% and 0.03% during year 4. These levels were accomplished during 1998 with one-half the density of dispensers per hectare, because this was the first year without a cost subsidy from the program. Also, secondary pests did not arise with the reduced insecticide pressure, as had initially been feared would occur. On the contrary, comparison plots under conventional practice experienced higher levels of secondary pests and often lower levels of beneficial insects and mites than did the CAMP program plots (Brunner et al., 2001). In the Parker Valley of Arizona with over 10 000 ha of cotton, growers in the late 1980s mandated an areawide mating disruption program due to this pest’s high level of resistance to insecticides (Staten et al., 1997). As in the codling moth CAMP program, damage diminished year by year in the Parker Valley with continued areawide application of pheromone mating disruption formulations. Whereas the percentage of infested bolls out of the tens of thousands that were sampled each year was more than 25% during mid-

August of year 1, during the same August period in year 2 (1990) damage was only 5.9% (Staten et al., 1997) (Table 21.2). At the mid-August point of year 3 damage was only 0.03% and by 1993, out of more than 22 000 bolls sampled seasonlong, not a single infested boll was found (0% damage) (Table 21.2). In contrast, the central Arizona average infestation rate for conventionally insecticide-treated hectarage in 1993 was 9–10% by mid-August (Staten et al., 1997).

21.4.2 Behavioral mechanisms underlying mating disruption success There is evidence from many studies that what has now been more aptly termed “competition” (Card´e, 1990; Card´e & Minks, 1995) but had previously been called “confusion” or “false trail-following,” does in fact occur in matingdisruption-treated fields, especially those receiving discrete point sources of dispensers such as hollow fibers or ropes. It is unknown, but we regard it as unlikely, that male moths in fields treated with sprayable microcapsules, creating a nearly uniform fog of pheromone from the closely spaced, weak point source emitters, would be subject to the competition mechanism. The majority of species that have been tested for their responses to uniform clouds of pheromone quickly habituate when released in the pheromone fog, and cease


upwind flight after only 1or 2 seconds, reverting to cross-wind casting within the cloud (Kennedy et al., 1981; Baker, 1985; Justus & Card´e, 2002). Card´e & Minks (1995) and Card´e et al. (1997) hypothesized that combinations of mechanisms will likely be operating in concert to various degrees, depending on the type of pheromone emission by the dispensers of a particular mating disruption formulation, to reduce males’ abilities to respond to pheromone plumes from their females. For dispensers acting in a competitive mode, when males are being “confused” and flying upwind in the plumes of synthetic pheromone emitted by hollow fibers (Haynes et al., 1986; Miller et al., 1990) or Shin-Etsu ropes (Card´e et al, 1997; Stelinski et al. 2004, 2005), they are receiving high-concentration contacts with the strands of pheromone in those plumes, and habituation of the olfactory pathways is occurring as a result of the male remaining in upwind flight and continuing to maintain contact with those strong pheromone strands (Baker et al., 1998). Habituation that results in reduced upwind flight in response to Shin-Etsu ropes has been observed in pink bollworm males by Card´e et al. (1997), and Stelinski et al. (2003a, b) have found evidence of long-lasting adaptation of the peripheral pheromone receptors in tortricids. Stelinski et al. (2004, 2005) documented that males of four tortricid orchard pest species are attracted in various degrees to individual rope dispensers in the field but often do not proceed all the way to the dispenser. Miller et al. (2006a, b) developed an intriguing new mathematical foundation for judging the degree to which competition (attraction) and non-competitive (habituation) mechanisms contribute to the efficacy of a particular mating disruptant formulation and dispenser deployment density. As formulations use increasingly widely spaced, high emission rate strategies for emitting pheromones, optimal attraction using the blends most closely approximating the natural female blend will become necessary to prolong the time a male spends locked onto the plume dosing himself with high amounts of pheromone in the plume strands to become habituated. At the other extreme, sprayable microcapsule formulations will likely depend very little,

if at all, on the attraction–competition mechanism, relying almost exclusively on habituation or plume camouflage (Card´e, 1990; Card´e & Minks, 1995). Regardless of the type of formulation, it is expected that all should be able to take advantage of the advanced period of sexual responsiveness that males exhibit before females begin to emit pheromone.

21.4.3 How mating disruption may suppress population growth It had routinely been assumed that successful mating disruption can only occur if the majority of females in a population are prevented from mating after the application of a mating disruption formulation. In reality, the females’ ability to obtain their first or second matings merely needs to be impaired and delayed. In all but a handful of the huge number of mating disruption field trials that have been conducted over the years, disruption of mating with freely flying females has not been directly assessed. The use of tethered females or clipped-wing females placed on “mating tables” or at stations deployed throughout the disruption plot and then dissected for the presence (mated) or absence (unmated) of spermatophores does not assess what is really happening to feral females that have the ability to fly freely throughout the area. The tethered female technique gives the illusion of being a robust, real-world assessment of mating disruption efficacy, and although it is one of many good indicators, it is especially deficient when the insects distribute themselves unevenly within the habitat due to environmental factors such as heat, humidity and wind. Unless the researcher knows ahead of time the locations where adults typically are most densely clumped and can tether the females there, the ability of the disruptant formulation to keep males from finding females will be overestimated. Knight (1997) introduced the concept of delayed mating of females within mating disruption plots, based on the relatively high proportion of codling moth females that he found had mated in mating disruption orchards, yet the formulation successfully reduced damage. Earlier studies on the oriental fruit moth that monitored the mating success of freely flying females had




suggested that something other than elimination of mating was operating. For the highly successful and grower-accepted Shin-Etsu rope formulation, Rice & Kirsch (1990) found that in plot after plot treated with mating disruptant, females’ abilities to mate at least once in disruption-treated plots was suppressed at most by 50% relative to check plots during a flight. The suppression of mating by the disruptant was often as little as 15–20%, despite the reduction of fruit damage in these plots to acceptable levels comparable to those using standard insecticide regimes (Rice & Kirsch, 1990). Thousands of females were captured in terpinyl acetate bait pails and analyzed for the presence or absence of spermatophores in their bursae copulatrices in both the disruptant-treated plots and the check plots. Vickers et al. (1985) and Vickers (1990) reported similar results for oriental fruit moth feral female mating success in Australian rope-treated mating disruption plots (23% mated females in mating disruption plots versus 90% in check plots). Delayed mating was directly confirmed in studies on European corn borer (Ostrinia nubilalis) using very high-release-rate metered semiochemical timed release system (MSTRSTM : dispensers (Fadamiro et al., 1999). Analyses were made of the bursae copulatrices of more than 2400 feral females that were captured by hand-netting during the daylight hours as they were flushed from their grassy aggregation areas. During each of the two summer flights, 100% of the females eventually became mated despite the application of high-release-rate, low-point-source density dispensers (Fadamiro et al., 1999). During the first flight about 50% of the females were virgin for the first few days of the flight in the mating disruption plots, but the mating success of females in these plots eventually reached 100% during the ensuing weeks as the flight proceeded. However, females attained this 100%-mated status more slowly in the mating disruption plots than in the check plots, in which females were 100% mated beginning at day 1. Analysis of the number of matings by (number of spermatophores in) European corn borer females showed that throughout the entire flight females captured in the disruption plots were attaining first and second matings at a significantly lower rate than those from the check

plots (Fadamiro et al., 1999). The mating disruptant was impairing the ability of females to attract and mate with males on a constant, daily basis but it did not completely eliminate mating. The application of this MSTRS formulation has subsequently been shown to reduce damage to corn by an average of 50–70% in various trials (T. C. Baker, unpublished data). Thus, as demonstrated in the oriental fruit moth (Rice & Kirsch, 1990) and codling moth studies (Knight, 1997), mating disruption success does not require keeping the population of females virgin, but rather just needs to impede females’ ability to attract males and retard the dates at which they achieve their first or even second matings. Retarding the dates at which first or second matings occur achieved significantly affects fecundity in the European corn borer and codling moth (Knight, 1997; Fadamiro and Baker, 1999).

21.4.4 Methods to assess the efficacy of mating disruption Of primary importance to growers and to companies marketing mating disruption products is the ability of a formulation to reduce crop damage to acceptable levels. In this context, “successful” mating disruption means assessing crop damage in pheromone-treated plots versus untreated check plots and finding that damage in mating disruption plots is lower. This process is problematic, but considered by many to be a significant outcome. The assessment is relatively straightforward, but it is essential that plots be large enough to reduce the probability that significant numbers of gravid females can fly in from nearby untreated plots and confound the damage data in the pheromonetreated plots. Being an indirect measure of actual mating disruption efficacy, damage assessment evaluates the end result of many processes that are of interest to those operating in commercial IPM and agronomic arenas. Conclusions that mating disruption was “successful” in the context of suppressing damage and being cost-effective can be arrived at without knowing exactly to what degree the formulation affected the behavior of males to reduce mating by freely flying feral females. Considered over many years to be the ultimate test of mating disruption efficacy, tethering


females either on a thread or by clipping their wings and placing them on open arenas so that they cannot move from the location at which they are placed has been a good tool in assessing mating disruption efficacy (see Evendon et al., 1999a, b), but this is only one of several indirect measures. If the adult moths in the natural population reside in a more clumped distribution in some preferred substructure of the vegetative habitat, placing females on artificial stations outside of each of these clumps will overestimate the efficacy of the disruption formulation with regard to preventing mating of feral females. The formulation in effect will only be assessed for its ability to prevent the tethered females from attracting males such that they leave their aggregation sites. This long-distance attraction will be easier to disrupt than will be the disruption of males within the same clumps that also contain females. Another limitation to this technique is that once a female mates with the first male arriving at her station, she emits no more pheromone and the ability to assess disruption of communication using that female ends. The data are binomial, with no opportunity for a graded assessment from individual females as to how many males the female could have attracted had it been calling for the entire activity period that night and not been mated. In another technique, a few virgin females are placed in small screen cages containing sugar water source to keep the females alive and hydrated (see Card´e et al., 1977). The cage is situated within a sticky trap, and so males that are attracted to the calling virgin females become ensnared before reaching the females. Males that do manage to land on the cage containing the females without getting trapped will still not be able to mate with any female in the cage. We feel that this technique has many advantages over other indirect measurement techniques such as the use of tethered females. First, as with tethered females, the caged females are emitting their natural blend at its natural emission rate. However, unlike tethered females, the caged females in traps provide a graded assessment of their ability to attract males. If males are able to get close enough to the calling females that they can be trapped, they certainly

will have mated if given that opportunity at such close range. The final step, mating, is an unnecessary one to examine because if the disruptant was not able to prevent a male’s long-distance orientation to the female’s plume, certainly it will not be sufficient to stop the male’s orientation to the female over the last 10–20 cm or so. The most widely used technique for assessing the disruption of mate-finding communication by disruptant formulations is the use of standard pheromone monitoring traps containing synthetic pheromone lures (see Rice & Kirsch, 1990; Knight, 1997; Staten et al., 1997; Baker et al., 1998; Fadamiro et al., 1999). This technique can be as informative as the use of caged calling females if the lure that is used has been shown previously in untreated check plots to be able to attract equivalent numbers of males as do caged calling females. The advantage of assessing trap capture reduction is that it provides a robust, graded data set from these continuously emitting sources. The disruption formulation is challenged throughout the attraction period each night or evening for its ability to continuously suppress the ability of males to locate “females.” Using monitoring traps baited with a good synthetic lure is also easier and less problematic than is the use of live, calling females.

21.4.5 Pheromone component blend composition in the disruptant formulation Minks & Card´e (1988) and Card´e & Minks (1995) reviewed results from many key sex pheromone communication disruption experiments and concluded that for a given species, the synthetic blend compositions and ratios most closely mimicking the natural blend for that species should be the most effective mating disruptants at a given dose per hectare because they will be able to make use of more of the mechanisms (see above) that result in disruption than will suboptimal, partial blends or off-ratios. Although the majority of field experimental evidence supports this conclusion, this does not mean that one or more of the more minor components, due to their more subtle effects, might be able to be eliminated from the final formulated product and still retain sufficient efficacy. There are, however, a few apparent




exceptions to this general rule (see Evendon et al., 1999a, b, c, 2000).

21.4.6 Some limitations to mating disruption A primary obstacle to the use of pheromone mating disruption on any crop is that it must be applied no later than the start of the first adult flight period, that is, before a grower knows for sure that there is even going to be a pest problem that season. The requirement of this “up-front” investment in pheromone puts mating disruption at a disadvantage compared to curative pest management tools that provide a wait-and-see option (see also Chapter 4). For example, insecticide applications can be made after a preliminary assessment of the population density during the first flight of adults or even after oviposition when larval damage can be assessed. Pheromone mating disruption cannot be used curatively in that manner. Charmillot (1990) summed up lessons he and his co-workers learned concerning when it is that mating disruption becomes less effective for codling moth population suppression. His lessons regarding codling moth are applicable to mating disruption efforts on other pests as well. First, mating disruption works best when applied on an areawide basis and is not advisable on very small areas (less than 1 ha). The crop borders represent vulnerable edges for immigration of mated females from adjoining untreated crop areas, and they also serve as a zone that concentrates females along the borders when there are no nearby planting of the same crop. Due to geometry, small hectarage accentuates the problem because the edge-to-area ratio becomes greater. Also, the use of more widely spaced dispensers requires that borders be given special attention to reduce the presence of pheromone-free clean-air “holes” along the borders. This can be accomplished either by decreasing the spacing between dispensers, or else through the use of different dispenser technology along borders, such as sprayable microcapsules. Except for a few small groups or pairs of species, pheromones are extremely speciesspecific. The replacement of broad-spectrum insecticides in some IPM systems with mating disruption that targets only one species in a pest com-

plex, might lead to increases in the populations of species that had been only secondary pests before mating disruption was used (see Rice & Kirsch, 1990). No pheromone disruption formulation has been created thus far that functions effectively as a “broad-spectrum” formulation. Evendon et al. (1999b, c) experimented with single blend formulations targeting obliquebanded leafroller (Choristoneura rosaceana) and threelined leafroller (Pandemis limitata) with some success. Perhaps there are other situations where effective multi-species blend formulations can be developed, although there are still none that have attained commercial success. Pheromone mating disruption formulations have to this point been expensive compared with curative applications of insecticides. Although growers of many crops are now well aware that mating disruption “works” as a pest management tool, pheromones’ expense plus the up-front nature of the cost has put them at a disadvantage relative to curative pest management tools. The active ingredient, the pheromone itself, is the most expensive part of a formulation. Costs per gram of even the least expensive pheromone components are approximately $US 1.00, and many other more expensive major pheromone components cost $US 3 to 20 per gram. Formulating the active ingredient into specialized dispensers adds to the cost, but nevertheless, if cheaper organic synthetic routes to the active ingredients can be developed, costs of the final formulated products could be reduced substantially.

21.5 Conclusions Insect pheromone-related technologies for monitoring endemic pest populations, detecting invasive species, mass trapping for population suppression and mating disruption have had a relatively recent history of development in IPM compared to biological control and insecticide technologies. New progress in the application of pheromones in IPM is being made in many areas, including the knowledge that mass trapping can be a highly effective and economically beneficial use of these behavior-modifying chemicals. Novel lure-and-trap technologies continue


to be developed for new pest species as they come on the scene in various regions of the world. New insights are also being made regarding ways to determine empirically the modes of action of mating disruption formulations, and the acceptance of the mating disruption technique by growers and government agencies has continued to grow in recent years. It remains to be seen whether other behavior-modifying chemicals such as host plant volatiles can become as widely used as pheromones for insect IPM in field situations where pheromones have been an integral part of insect IPM programs for approximately 35 years.

References Agnello, A. M., Kovach, J., Nyrop, J. P. et al. (1994). Extension and evaluation of a simplified monitoring program in New York apples. American Entomologist, 40, 37–49. Baker, T. C. (1985). Chemical control of behaviour. In Comprehensive Insect Physiology, Biochemistry, and Pharmacology, eds. G. A. Kerkut & L. S. Gilbert, pp. 621–672. Oxford, UK: Pergamon Press. Baker, T. C., Fadamiro, H. Y. & Coss´e, A. A. (1998). Moth uses fine tuning for odor resolution. Nature, 393, 530. Baker, T. C., Staten, R. T. & Flint, H. M. (1990). Use of pink bollworm pheromone in the southwestern United States. In Behavior-Modifying Chemicals for Insect Management, eds. R. L. Ridgway, R. M. Silverstein & M. N. Inscoe, pp. 417–436. New York: Marcel Dekker. Brunner, J., Welter, S., Calkins, C. et al. (2001). Mating disruption of codling moth: a perspective from the Western United States. IOBC/WPRS Bulletin, 25, 207– 215. Butenandt, A. (1959). Wirkstoffe des Insektenveiches. Naturwissenschaften, 46, 461–471. Card´e, R. T. (1990). Principles of mating disruption. In Behavior-Modifying Chemicals for Insect Management, eds. R. L. Ridgway, R. M. Silverstein & M. N. Inscoe, pp. 47–71. New York: Marcel Dekker. Card´e, R. T. & Minks, A. K. (1995). Control of moth pests by mating disruption: successes and constraints. Annual Review of Entomology, 40, 559–585. Card´e, R. T., Baker, T. C. & Castrovillo, P. J. (1977). Disruption of sexual communication in Laspeyresia pomonella (codling moth), Grapholitha molesta (oriental fruit moth) and G. prunivora (lesser appleworm) with hollow fiber attractant sources. Entomologia Experimentalis et Applicata 22, 280–288.

Card´e, R. T., Mafra-Neto, A., Staten, R. T. & Kuenen, L. P. S. (1997). Understanding mating disruption in the pink bollworm moth. IOBC/WPRS Bulletin, 20, 191–201. Charmillot, P. J. (1990). Mating disruption technique to control codling moth in western Switzerland. In Behavior-Modifying Chemicals for Insect Management, eds. R. L Ridgway, R. M. Silverstein & M. N. Inscoe, pp. 165– 182. New York: Marcel Dekker. Chinchilla, C. M. & Oehlschlager, A. C. (1992). Capture of Rhynchophorus palmarum in traps baited with the male-produced aggregation pheromone. ASD Oil Palm Papers, 5, 1–8. Elkinton, J. S. & Card´e, R. T. (1981). The use of pheromone traps to monitor distribution and population trends of the gypsy moth. In Management of Insect Pests with Semiochemicals, ed. E. R Mitchell, pp. 41–55. New York: Plenum Press. Evendon, M. L., Judd, G. J. R. & Borden, J. H. (1999a). Pheromone-mediated mating disruption of Choristoneura rosaceana: is the most attractive blend really the most effective? Entomologia Experimentalis et Applicata, 90, 37–47. Evendon, M. L., Judd, G. J. R. & Borden, J. H. (1999b). Mating disruption of two sympatric, orchard-inhabiting tortricids, Choristoneura rosaceana and Pandemis limitata (Lepidoptera: Tortricidae), with pheromone components of both species’ blends. Journal of Economic Entomology, 92, 380–390. Evendon, M. L., Judd, G. J. R. & Borden, J. H. (1999c). Simultaneous disruption of pheromone communication in Choristoneura rosaceana and Pandemis limitata with pheromone and antagonist blends. Journal of Chemical Ecology, 25, 501–517. Evendon, M. L., Judd, G. J. R. & Borden, J. H. (2000). Investigations of mechanisms of pheromone communication disruption of Choristoneura rosaceana (Harris) in a wind tunnel. Journal of Insect Behavior, 13, 499– 510. Fadamiro, H. Y. & Baker, T. C. (1999). Reproductive performance and longevity of female European corn borer, Ostrinia nubilalis: effects of multiple mating, delay in mating, and adult feeding. Journal of Insect Physiology, 45, 385–392. Fadamiro, H. Y., Coss´e, A. A. & Baker, T. C. (1999). Mating disruption of European corn borer, Ostrinia nubilalis, by using two types of sex pheromone dispensers deployed in grassy aggregation sites in Iowa cornfields. Journal of Asia-Pacific Entomology, 2, 121– 132. Hardee, D. D., Mitchell, E. B. & Huddleston, P. M. (1967a). Field studies of sex attraction in the boll weevil. Journal of Economic Entomology, 60, 1221–1224.




Hardee, D. D., Mitchell, E. B. & Huddleston, P. M. (1967b). Procedure for bioassaying the sex attractant of the boll weevil. Journal of Economic Entomology, 60, 169– 171. Haynes, K. F., Li, W. G. & Baker, T. C. (1986). Control of pink bollworm moth (Lepidoptera: Gelechiidae) with insecticides and pheromones (attracticide): lethal and sublethal effects. Journal of Economic Entomology, 79, 1466–1471. Justus, K. A. & Card´e, R. T. (2002). Flight behaviour of two moths, Cadra cautella and Pectinophora gossypiella, in homogeneous clouds of pheromone. Physiological Entomology, 27, 67–75. Kennedy, J. S., Ludlow, A. R. & Sanders, C. J. (1981). Guidance of flying male moths by wind-borne sex pheromone. Physiological Entomology, 6, 395– 412. Knight, A. L. (1997). Delay of mating of codling moth in pheromone disrupted orchards. IOBC/WPRS Bulletin, 20, 203–206. Madsen, H. F. (1981). Monitoring codling moth populations in British Columbia apple orchards. In Management of Insect Pests with Semiochemicals, ed. E. R. Mitchell, pp. 57–62. New York: Plenum Press. Miller, E., Staten, R. T., Nowell, C. & Gourd, J. (1990). Pink bollworm (Lepidoptera: Gelechiidae): point source density and its relationship to efficacy in attracticide formulations of gossyplure. Journal of Economic Entomology, 83, 1321–1325. Miller, J. R., Gut, L. J., de Lame, F. M. & Stelinski, L. L. (2006a). Differentiation of competitive vs. noncompetitive mechansisms mediating disruption of moth sexual communication by point sources of sex pheromone. I. Theory. Journal of Chemical Ecology, 32, 2089–2114. Miller, J. R., Gut, L. J., de Lame, F. M. & Stelinski, L. L. (2006b). Differentiation of competitive vs. noncompetitive mechanisms mediating disruption of moth sexual communication by point sources of sex pheromone. II. Case studies. Journal of Chemical Ecology, 32, 2115–2144. Minks, A. K. & Card´e, R. T. (1988). Disruption of pheromone communication in moths: is the natural blend really most efficacious? Entomologia Experimentalis et Applicata, 49, 25–36. Oehlschlager, A.C., Pierce, H.D., Morgan, B. et al. (1992). Chirality and field testing of Rhynchophorol, the aggregation pheromone of the American palm weevil. Naturwissenschaften (Berlin), 79, 134– 135. Oehlschlager, A. C., Chinchilla, C. M., Gonzales, L. M. et al. (1993). Development of a pheromone-based trapping

system for Rhynchophorus palmarum (Coleoptera: Curculionidae). Journal of Economic Entomology, 86, 1381– 1392. Oehlschlager, A. C., Chinchilla, C., Castillo, G. & Gonzalez, L. (2002). Control of red ring disease by mass trapping of Rhynchophorus palmarum (Coleoptera: Curculionidae). Florida Entomologist, 85, 507– 513. Rice, R. E. & Kirsch, P. (1990). Mating disruption of oriental fruit moth in the United States. In BehaviorModifying Chemicals for Insect Management, eds. R. L. Ridgway, R. M. Silverstein & M. N. Inscoe, pp. 193–211. New York: Marcel Dekker. Ridgway, R. L., Inscoe, M. N. & Dickerson, W.A. 1990. Role of the boll weevil pheromone in pest management. In Behavior-Modifying Chemicals for Insect Management, eds. R. L. Ridgway, R. M. Silverstein & M.N. Inscoe, pp. 437– 471. New York: Marcel Dekker. Riedl, H. & Croft, B. A. (1974). A study of pheromone trap catches in relation to codling moth (Lepidoptera: Olethreutidae) damage. Canadian Entomologist, 112, 655–663. Riedl, H., Croft, B. A. & Howitt, A. G. (1976). Forecasting codling moth phenology based on pheromone trap catches and physiological-time models. Canadian Entomologist, 108, 449–460. Showers, W. B., Smelser, R. B., Keaster, A. J. et al. (1989a). Recapture of marked black cutworm (Lepidoptera: Noactuidae) males after long-range transport. Environmental Entomology, 18, 447–458. Showers, W. B., Whitford, F., Smelser, R. B. et al. (1989b). Direct evidence for meteorologically driven longrange dispersals of an economically important moth. Ecology, 70, 987–992. Staten, R. T., El-Lissy, O. & Antilla, L. (1997). Successful area-wide program to control pink bollworm by mating disruption. In Insect Pheromone Research: New Directions, eds. R. T. Card´e & A. K. Minks, pp. 383–396. New York: Chapman & Hall. Stelinski, L. L., Miller, J. R. & Gut, L. J. (2003a). Presence of long-lasting peripheral adaptation in the obliquebanded leafroller, Choristoneura rosaceana and the absence of such adaptation in the redbanded leafroller, Argyrotaenia velutinana. Journal of Chemical Ecology, 29, 405–423. Stelinski, L. L., Gut, L. J. & Miller, J. R (2003b). Concentration of air-borne pheromone required for long-lasting peripheral adaptation in the obliquebanded leafroller, Choristoneura rosaceana. Physiological Entomology, 28, 97–107. Stelinski, L. L., Gut, L J., Pierzchala, A. V. & Miller, J. R. (2004). Field observations quantifying attraction of


four tortricid moth species to high-dosage pheromone rope dispensers in untreated and pheromone-treated apple orchards. Entomologia Experimentalis et Applicata, 113, 187–196. Stelinski, L. L., Gut, L. J., Epstein, D. & Miller, J. R. (2005). Attraction of four tortricid moth species to high dosage pheromone rope dispensers: observations implicating false plume following as an important factor in mating disruption. IOBC/WPRS Bulletin, 28, 313–317. Tumlinson, J. H., Hardee, D. D., Minyard, J. P. et al. (1968). Boll weevil sex attractant: isolation studies. Journal of Economic Entomology, 61, 470–474. Tumlinson, J. H., Hardee, D. D., Gueldner, R. C. et al. (1969). Sex pheromones produced by male boll weevil: Isolation, identification, and synthesis. Science, 166, 1010–1012.

Tumlinson, J. H., Gueldner, R. C., Hardee, D. D. et al. (1970). The boll weevil sex attractant. In Chemicals Controlling Insect Behavior, ed. M. Beroza, pp. 41–59. New York: Academic Press. Tumlinson, J. H., Gueldner, R. C., Hardee, D. D. et al. (1971). Identification and synthesis of the four compounds comprising the boll weevil sex attractant. Journal of Organic Chemistry, 36, 2616–2621. Vickers, R. A. (1990). Oriental fruit moth in Australia and Canada. In Behavior-Modifying Chemicals for Insect Management, eds. R. L. Ridgway, R. M. Silverstein & M. N. Inscoe, pp. 183–192. New York: Marcel Dekker. Vickers, R. A., Rothschild, G. H. L. & Jones, E. L. (1985). Control of the oriental fruit moth, Cydia molesta (Busck) (Lepidoptera:Tortricidae), at a district level by mating disruption with synthetic female pheromone. Bulletin of Entomological Research, 75, 625–634.


Chapter 22

Insect endocrinology and hormone-based pest control products in IPM Daniel Doucet, Michel Cusson and Arthur Retnakaran IPM methods were developed largely in response to the negative consequences of the intensive use of broad-spectrum pesticides in the early to mid twentieth century (Kogan, 1998). These insecticides, belonging to the carbamate, organophosphate and organochlorine families, have unintended side effects such as environmental persistence, bioaccumulation, development of resistance among target pests, toxicity to non-target species (especially natural enemies) and human health risks. While IPM focuses mainly on preventative tactics (e.g. crop rotation) rather than remedial ones, synthetic chemical insecticides are still very much needed to achieve effective control in many agricultural systems. The study of insect physiology has been driven, in no small part, by the need for safe alternatives to broad-spectrum insecticides. Theoretically at least, digestion, excretion, neuronal communication, metabolism and other physiological processes all comprise “insect-specific” components that are vulnerable and could be targeted by synthetic molecules. To this day, however, IPMcompatible pest control products that target the insect endocrine system far outnumber those targeting other systems. In particular, hormone mimics that control development have enjoyed not only wide appeal but also many commercial successes, and additional control products targeting hormone production and function are currently

under development. In this chapter we provide an overview of (1) insect endocrinology, (2) existing control products that mimic ecdysone and juvenile hormone (JH) action and (3) possible development of disruption control strategies based on novel endocrine functions that are likely to generate new IPM tools in the future.

22.1 Basics of hormone biochemistry and biology The term “hormone” was first coined by Ernest Starling a hundred years ago to define any chemical messenger, secreted by an organ, which travels through the bloodstream to affect the physiology of another, distant organ (reviewed in Henderson, 2005). At that time, hormones were known to be important mediators of vertebrate physiology, but Kopec (1917) soon demonstrated that similar secretions existed in insects, with his isolation of a brain factor promoting molting in a lepidopteran larva. Insect hormones fall into four classes based on their chemical structures: (1) peptide and protein hormones, (2) biogenic amines, (3) prostaglandins and (4) terpenoid lipids. Peptide hormones are chains of amino acids usually shorter than 20–30 residues. Longer chains are traditionally referred

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


to as proteins. This is by far the most abundant class of insect hormones, with several hundred different peptides isolated to date, from various species. Biogenic amines are derivatives of amino acids and are involved in signal transduction. In insects, octopamine and tyramine are two such compounds that regulate important aspects of locomotor and non-locomotor behavior, circadian rhythms and stress response (Gole & Downer, 1979; Fussnecker et al., 2006). Prostaglandins are oxygenated metabolites of arachidonic acid and play diverse roles including mediation of cellular immunity and release of egg-laying behavior (Stanley, 2006). Terpenoid hormones are lipid molecules constructed from the basic hydrocarbon unit 2-methyl-1,3-butadiene, also called isoprene. In insects, two important subclasses exist: the juvenile hormones, which are derivatives of linear chains of three isoprene units (i.e. sesquiterpenes) and the ecdysteroids, which consist of a tetracyclic cholestane ring system derived from cholesterol, which is elaborated from smaller isoprenoids. Both hormones act on the timing and nature of molting in all insects. The most successful hormone-based insecticides either mimic or inhibit the activity of these hormones and are presented in greater detail in Sections 22.4 and 22.5.

22.2 Overview of the insect endocrine system Hormone secretion in insects is accomplished by a limited number of glands, organs and tissues, and, as observed in other animal groups, the central nervous system (CNS) plays an overarching role. The CNS integrates sensory information and translates it into nervous or hormonal outputs that bring about the physiological, behavioral and developmental processes necessary for survival and reproduction (Nijhout, 1994). The endocrine control of these processes by the CNS can be either direct or indirect. A good example of direct control is the induction of diapause in silkworm (Bombyx mori) by the diapause hormone (DH), released by secretory neurons from the subesophageal ganglia (Hasegawa, 1957; Sato et al.,

1993). More frequently, however, the CNS modulates the secretions of other endocrine glands by releasing inhibitory or stimulatory (tropic) neurohormones. This hierarchical organization between hormones and neurohormones can be highly complex and include endocrine feedback loops. Such is the case for insect ecdysis, where the proper unfolding of this innate behavior is ensured by a tightly controlled spatial and temporal release of at least half a dozen hormones (Kim et al., 2006). While a thorough review of insect hormones (and their sites of production) would be beyond the scope of this chapter, a survey of some of the more important ones was presented by Doucet et al. (2007a).

22.2.1 CNS and associated neurohemal organs The CNS is composed of two major endocrine centers: the brain–retrocerebral complex (RC) and the perisympathetic organs (PSOs) of the ventral nerve cord. The RC is a bipartite structure, posterior to the brain, composed of the corpora allata (CA) and corpora cardiaca (CC). Clusters of neurosecretory cells are located in various parts of the insect brain (e.g. medial, lateral and ventral) and the majority release their secretions via the CC (Nijhout, 1994). Some brain neurosecretory cells instead use the CA as a neurohemal organ (e.g. for the secretion of prothoracicotropic hormone, PTTH), while fewer still release their secretions distally, for example in the vicinity of the proctodeum (hindgut) (e.g. the proctodeal nerves in tobacco hornworm [Manduca sexta] are the release sites of eclosion hormone [EH], which is synthesized by the ventromedial neurosecretory cells in the brain (Truman & Copenhaver, 1989). The glandular portions of the CA and the CC also secrete their own hormones besides those from the neurosecretory cells: the CA secrete juvenile hormone (JH) (see Section 22.4) while the CC contain endocrine cells that produce peptides such as the adipokinetic hormone (AKH). PSOs are segmentally distributed neurohemal organs that are functionally close to the CC (Nijhout, 1994). However, recent mass spectrometry analysis of PSO extracts indicate that they release a cocktail of neuropeptides distinct from those produced by the CC (Predel et al., 1999, 2000).




22.2.2 Other endocrine glands While the CNS is the most structurally and functionally complex hormone production site in insects, non-neural peripheral secretory organs contribute in important ways to the regulation of physiological processes. For example, a few highly specialized glands such as the prothoracic glands (PGs), the Inka cells in the epitracheal glands and the epiproctodeal glands (EPGs) have confirmed or suspected roles in endocrine regulation of molting. PGs are located in the prothoracic segments of immature insects and are devoted to the secretion of the hormone ecdysone. The Inka cells come into play later during the molting cycle, at the time of ecdysis. These cells are specialized in the secretion of two peptides regulating ecdysis-related behaviors: the pre-ecdysis triggering hormone (PETH) ˇ nan and the ecdysis triggering hormone (ETH) (Zitˇ et al., 2003). The number, morphology and distribution of Inka cells within the body can be quite variable among insects of different orders, but they are always in close association with the epitracheal glands, near the spiracles. EPGs are secretory structures that have been described in Manduca. They are present as a pair of multinucleated cells located at the junction of the hindgut and the rectum, in close contact with the proctodeal nerve (Davis et al., 2003). EPGs synthesize a myoinhibitory-like peptide (MIP-like I) that is speculated to shut off ecdysone production by the PGs, at the end of each molt (Davis et al., 2003). Non-specialized endocrine tissues include the gut, the fat body and ovaries. Midgut endocrine cells have features of typical secretory cells (e.g. abundant secretory granules, clear cytoplasm: Brown et al., 1985; Neves et al., 2003) and have been found to release peptides involved in midgut contraction and other aspects of digestive physiology. Hormones secreted by these cells include allatostatins and allatostatin-like peptides (ASlike) (Davey et al., 2005) and crustacean cardioactive peptide (CCAP) (Sakai et al., 2004). Interestingly, both AS and CCAP were originally discovered in functions unrelated to invertebrate midgut physiology: as an inhibitor of JH synthesis by the CA (Stay & Tobe, 2007) and CCAP as a peptide regulating heartbeat in the crab and ecdysis in ˇ nan & Adams, 2005). Fat body cells from insects (Zitˇ the desert locust (Schistocerca gregaria) have been

shown to synthesize neuroparsins (Claeys et al., 2003). Very little is known about the function of these neuroparsins, but their expression is stageand sex-dependent, and is regulated by JH and 20-hydroxyecdysone (Claeys et al., 2006). Rachinsky et al. (2006) also detected low levels of allatotropin, a JH biosynthesis stimulating peptide, in the fat body of Manduca sexta. Finally, the ovaries of many insects constitute a source of ecdysteroids that play important roles in reproductive physiology. These steroids either stimulate the transcription of yolk protein precursor genes in females (e.g. vitellogenin, lipophorin: Raikhel et al., 2005) or presumably assist in cuticle formation during the development of the embryo (Lagueux et al., 1984).

22.3 Ecdysone and ecdysone agonists 22.3.1 Ecdysone functions The elucidation of the structure of ecdysone (Butenandt & Karlson, 1954) was the major catalyst for research on the physiology of the molting hormone and its role in insect metamorphosis. We now know that the biologically active form of ecdysone is 20-hydroxyecdysone (20E), which acts as the ligand of a heterodimeric receptor system consisting of two nuclear receptors, the ecdysone receptor (EcR) and ultraspiracle (USP). USP is an allosteric effector for ligand binding by the EcR. 20E binds to the ligand-binding domain of the EcR subunit of the EcR–USP dimer (EcR complex). In turn, this complex binds to a specific sequence on the DNA, designated as the “ecdysone response element” (EcRE), located upstream from the gene that is to be activated or repressed (Nordeen et al., 1998). Binding of the 20E-liganded EcR complex to an EcRE causes transactivation of gene transcription. Conversely, the unliganded complex causes repression. The degree of binding affinity appears to correlate with activity. The bipolar transcriptional ability is probably aided by corepressors and coactivators, as in vertebrate steroid hormone receptor systems, but details are lacking at present (King-Jones & Thummel, 2005).


22.3.2 Molting regulation by 20E The major function of 20E is the regulation of the molting process in insects. In the Lepidoptera, this hormone appears as a single peak during the middle of each larval stadium, except for the last one where there are two peaks; the earlier, smaller peak is the commitment peak which triggers the reprogramming towards pupal development, followed by the larger molt peak. The presence of high titers of JH during the larval 20E peak defines the molt as larval whereas the absence of JH during the commitment peak results in pupal transformation, which involves the turning off of larval genes and turning on of pupal genes. The physiological and molecular basis of the switch has been elegantly elucidated by Riddiford and her associates showing for instance that the Broad Complex gene (BR-C) encodes the transcription factor that specifies the pupal cuticle (Riddiford et al., 2003). Thus, 20E can both repress and activate genes, and its effects can be modulated by JH.

22.3.3 Steroidal and non-steroidal analogs of ecdysone Various plant ecdysteroids (phytoecdysones) and synthetic steroids have been assessed for their ability to act as agonists or antagonists of ecdysone and for their insecticidal activity, but with limited success. Besides being difficult to synthesize, steroids are large molecules lacking contact activity and are prone to oxidative breakdown. Over 200 plant ecdysteroids have been studied and none of them has shown high potential as control products (Dinan, 2001). Similarly, several synthetic steroids were tested for biological activity, but none proved very effective (Robbins et al., 1970). A few steroid analogs such as cucurbitacins and brassinosteroids were tested for antagonistic activity but, again, the effects were weak (Charrois et al., 1996; Dinan et al., 1997). More recently, however, non-steroidal ecdysone agonists such as tetrahydroquinoline compounds showed some activity, especially against mosquitoes (Palli et al., 2005). By far the most effective compounds among the non-steroidal ecdysone agonists are the diacylhydrazines, serendipitously discovered in the Rohm and Haas (RH) laboratories at Spring House, PA (Hsu, 1991). Initial optimization following discovery of the chemistry led to a first promising

candidate, RH-5849, but soon three other compounds displaying greater activity were synthesized: (1) RH-5992 (tebufenozide) was active against several lepidopterans and was registered under the commercial names MimicR , ConfirmR and RomdanR ; (2) RH-2485 (methoxyfenozide) was several times more active than RH-5992 on lepidopterans, and was marketed under the names IntrepidR , RunnerR , ProdigyR and FalconR ; (3) RH-0345 or halofenozide was active against coleopterans in addition to lepidopterans and was registered under the name Mach 2R . Rohm and Haas since sold all these compounds to Dow Agrosciences, Indianapolis, IN. More recently Nippon Kayaku, Saitane and Sankyo, Ibaraki, from Japan, have come up with a new diacylhydrazine which they have named ANS-118/CM-001 or chromafenozide, which is active against lepidopterans and is registered under the names MatricR and KillatR (Nakagawa, 2005) (Fig. 22.1). Diacylhydrazines are functionally similar to 20E and bind to the EcR receptor complex as ligands (Nakagawa, 2005). While they mimic the natural hormone in triggering the molting process, the latter is never completed. Treated larvae show all the initial signs of molting, such as feeding cessation, head-capsule slippage (with the new head capsule remaining untanned), and loosening of the cuticle due to apolysis. However, the similarity between 20E action and diacylhydrazines stops at this juncture. There is no ecdysis, resumption of feeding, or sclerotization and darkening of the new head capsule that normally follow clearance of 20E. Rather, the larva is in a state of developmental arrest or suspended animation and eventually dies of starvation and desiccation.

22.3.4 Diacylhydrazines as pest-control products The non-steroidal ecdysone agonists have been tested on a variety of insects with varying degrees of success. The activity of different diacylhydrazines on a representative list of insects was presented by Doucet et al. (2007b). The MimicR formulation of RH-5992 (tebufenozide) works very well against the spruce budworm, a serious pest of the balsam fir (Abies balsamea) and white spruce (Picea glauca) in the boreal forests of North America (Retnakaran et al., 1997; Cadogan et al., 1998, 2005).




Fig. 22.1 Chemical structures of insect molting hormones, a phytoecdysteroid and five of the most promising diacylhydrazine ecdysone agonists.

Laboratory studies show that methoxyfenozide is about ten times more active than tebufenozide (Sundaram et al., 1998). The grape berrymoth (Lobesia botrana) is very sensitive to methoxyfenozide, which shows excellent potential for control of this pest. Older larvae are more susceptible than younger ones, and treated adults display reduced fecundity and fertility (S´ aenz-de-Cabez´ on Irigaray et al., 2005). In some instances the molt-inducing activity of diacylhydrazines is very weak, and thus this class of compounds cannot be retained as an effective control solution. The obliquebanded leafroller (Choristoneura rosaceana) shows relatively low susceptibility to tebufenozide and methoxyfenozide (Ahmad et al., 2002). In the case of the whitemarked tussock moth (Orgyia leucostigma), treatment with tebufenozide induces head capsule slippage but the larva, after remaining quiescent for a week to 10 days, becomes active and molts into the next instar (Retnakaran et al., 2003). In vitro and in vivo data tend to indicate that the potency

of diacylhydrazines is related to absorption and excretion rates in a given species (Retnakaran et al., 2001; Smagghe et al., 2001). Other cases of low susceptibility include the codling moth (Cydia pomonella) (Sauphanor & Bouvier, 1995), and the green-headed leafroller (Planotortrix octo) from New Zealand (Wearing, 1998), to name a few. This only goes to show that these compounds, while very effective on some species, have their own limitations and are not a general cure-all.

22.3.5 Ecdysone agonists as candidates for IPM Over the years pest management methods have progressively evolved towards an ecologically based systems approach, combining biological, cultural, physical and chemical tools in a way that minimizes economic, health and environmental risks. Such an IPM approach is a dynamic process constantly aiming at maximizing target specificity and minimizing environmental side effects. Ecdysone agonists have so far been shown to be insect-specific and, among the various agonists, some are far more active on one group of insects than others. In general, their mammalian toxicity is very low; the acute oral toxicity for


Fig. 22.2 Structures of juvenile hormones (JHs) produced by insects and of four JH analogs commercially available.

rat and mouse is >5000 mg/kg for tebufenozide, methoxyfenozide and chromafenozide, and >2850 mg/kg for halofenozide. These compounds have no detectable effects on reproduction, and are negative in the Ames mutation assay (Dhadialla et al., 2005). These non-steroidal ecdysone agonists are also relatively safe for the environment and do not have any negative impact on forest-litter-dwelling organisms such as earthworms and Collembola (Addison, 1996). Macroinvertebrates in freshwater ponds are unaffected by tebufenozide (Kreutzweiser et al., 1994). Tebufenozide and methoxyfenozide were shown to have little effect on the bumblebee (Bombus terrestris) (Mommaerts et al., 2006). A summary of safety to parasites, predators and pollinators is provided by the National Registration Authority of Australia (Anonymous, 2002). Tebufenozide was also found to have little impact on a generalist predator, the common green lacewing (Chrysoperla carnea) (Medina et al., 2003). The persistence, breakdown and catabolism of tebufenozide have been well studied. In conifer needles and forest litter some persistence was observed, but well within tolerance levels (Sundaram et al., 1996). Thus the environmental safety and narrow spectrum of activity of these non-steroidal ecdysone agonists make them a valuable addition to the arsenal of candidates for IPM.

22.4 Juvenile hormones 22.4.1 Chemical nature, biosynthesis and functions The juvenile hormones (JHs) form a family of lipophilic, sesquiterpenoid molecules with epoxide and methyl-ester functionalities. All are derived from the mevalonate pathway intermediate, farnesyl diphosphate (FPP), or from one of its ethyl-branched homologs. The majority of insects produce only one chemical form of JH, JH III (C-16), but the Lepidoptera produce four additional, ethyl-substituted JHs (JH 0 [C-19], JH I [C18], JH II [C-17] and 4-methyl JH I [C-19]) and the Diptera produce a bis-epoxy form of JH III (JH-B3) (Fig. 22.2). These hormones are all produced de novo by the CA. JH biosynthesis begins with the condensation of three units of acetyl-CoA (for JH III and JH-B3) or two units of acetyl-CoA and one of propionyl CoA (lepidopteran JHs), leading to the formation of the isoprene (C-5) and homo-isoprene (C-6) building blocks of FPP and ethyl-substituted FPPs. Unlike the enzymatic steps that are responsible for FPP formation, which are common to most living organisms, those that convert FPP into JH are specific to insects. Although JH plays multiple roles in insects, it owes its name to its juvenilizing effects during larval molts: the presence of elevated JH titers during ecdysone secretion represses the expression of metamorphic genes, thus maintaining the




insect in a juvenile (larval) state. Shortly after the final larval molt, JH biosynthesis ceases and a JHspecific catabolic enzyme known as JH esterase (JHE) is secreted, which leads to a rapid decline in the JH titer. Under these conditions, ecdysone triggers the metamorphic program leading to the pupal molt. In adult insects, JH is a gonadotropin and is best known for its stimulatory effects on vitellogenesis, inducing the production of vitellogenin by the fat body and/or its uptake by developing oocytes. In males, JH has been implicated in the production of accessory sex gland secretions and in control of courtship behavior. JH has also been shown to be involved in regulation of various other processes, including embryogenesis, migration, cast differentiation, polyphenism and reproductive diapause (Cusson, 2004). Although many of the roles played by JH have been well characterized, its mode of action at the molecular level remains unclear. Several proteins have been tentatively identified as JH receptors, but conclusive evidence regarding their role as receptors is still lacking (see Palli & Cusson, 2007, for a more complete account of recent work in this area). The presence of a JH response element (JHRE) in the promoter region of the JH-responsive JHE gene suggests that JH can act through a nuclear receptor. Some nuclear proteins do, indeed, bind to this JHRE, but binding requires that the proteins be dephosphorylated, a process that appears to be induced by JH (Kethidi et al., 2006).

22.4.2 Juvenile hormone analogs and IPM Following the initial isolation of JH, Caroll Williams (1956) predicted the dawn of JH-based insecticides. It was surmised that synthetic JH-like molecules would fatally interfere with JH functions and that insects would not likely develop resistance against such compounds (Williams, 1967). Many JH analogs (molecules with JH effects, with or without a JH-like terpenoid structure) were subsequently designed, synthesized and assayed for insecticidal activity (Sl´ ama et al., 1974). Some of them (e.g. methoprene, hydroprene, fenoxycarb, pyriproxifen, diofenolan; Fig. 22.2) were found to be effective against certain pests and have since enjoyed commercial success, particularly for the control of insects that are injurious in the adult stage (e.g. mosquitoes, fleas, whiteflies, etc.), but their efficacy against phytophagous larval insects

(e.g. caterpillars) has often proven to be limited, largely because the analogs interfere with metamorphosis, once larval feeding has ended (see Dhadialla et al., 1998, 2005 for reviews).

22.4.3 Anti-JHs It has long been recognized that a strategy involving the induction of precocious metamorphosis through the inhibition of JH biosynthesis would be better suited to the control of immature phytophagous insects than one involving the disruption of metamorphosis with JH analogs (Cusson & Palli, 2000). Although a number of naturally occurring and synthetic inhibitors of JH biosynthesis have been evaluated for their ability to trigger precocious metamorphosis, none has yet been developed commercially. The “precocenes,” isolated from the bedding plant (Ageratum houstonianum) originally generated much hope and enthusiasm as they caused premature metamorphosis in the milkweed bug (Oncopeltus fasciatus). However, these compounds proved to be ineffective against holometabolous insects. Inhibitors of mevalonate pathway enzymes known to have hypocholesterolemic activity in mammals were also tested on insects; examples include the fungal metabolite compactin, an inhibitor of HMGCoA reductase (Monger et al., 1982), and the fluorinated mevalonate analog, fluoromevalonate, an inhibitor of enzymes involved in the processing of mevalonate (Quistad et al., 1981), both of which were found to induce precocious metamorphosis in lepidopteran larvae, but required high doses and/or repeated applications. Allylic alcohol derivatives of dimethylallyl diphosphate, the C-5 chain initiator for FPP production by FPP synthase (FPPS), provided similar results (Quistad et al., 1985). Inhibitors of later, insect-specific steps of JH biosynthesis, such as formation of the methyl ester moiety and epoxidation, were also examined for their ability to block JH biosynthesis and induce precocious metamorphosis. Brevioxime, a compound with a sesquiterpene-like structure isolated from the fungus Penicillium brevicompactum, was shown to inhibit in vitro JH biosynthesis by the CA of migratory locus (Locusta migratoria) (Castillo et al., 1998). Similarly, the synthetic 1,5-disubstituted imidazole, KK-42, an inhibitor of the P450-linked enzyme that epoxidizes the JH precursor methyl farnesoate (MF), inhibited


in vitro JH biosynthesis by the CA of the migratory locust (Castillo et al., 1998) and the Pacific beetle cockroach (Diploptera punctata) (Pratt et al., 1990). Again, concentrations of the compounds required to achieve inhibition were relatively high. Recent developments in the area of insect genomics have led to the cloning and characterization of enzymes specific to JH biosynthesis, including JH acid methyl transferase from the silkworm (Shinoda & Itoyama, 2003) and MF epoxydase from Pacific beetle cockroach (Helvig et al., 2004). The ability to produce these enzymes in heterologous expression systems should lead to substantial improvements in our capacity to assess their three-dimensional structures and design potent and highly specific inhibitors that could be used as anti-JH insecticides. In addition to the above proteins, several mevalonate pathwayspecific enzymes have been cloned from various species of insects. Although these may not provide the most suitable target sites for insecticide development, given that they are found in most living organisms, the lepidopteran homolog of at least one of them, FPPS, appears to display significant structural singularities believed to be instrumental in the biosynthesis of the ethyl-substituted JHs (Cusson et al., 2006; Sen et al., 2007); this enzyme may thus prove to be a suitable target for the design of lepidoptera-specific inhibitiors (see Palli & Cusson, 2007 for a more in-depth review). Identification of a JH receptor is perhaps the most promising avenue for the development of highly effective anti-JH compounds. As indicated above, however, this receptor has so far remained elusive, despite sustained and continuing research efforts. Once a receptor has been isolated, it should become possible to design cellbased high-throughput assays for the screening of potential JH antagonists that are effective at the target tissue level (Minakuchi & Riddiford, 2006; Palli & Cusson, 2007).

22.5 Conclusion Insecticides targeting endocrine functions have a young history, and it remains to be seen at what rate new products will be successfully brought on the market. Agonists and antagonists of peptidic

hormones are currently being investigated, such as cyclic backbone peptides inhibiting the action of pheromone biosynthesis activating neuropeptide (PBAN) (Altstein, 2004). Other suitable targets for peptidomimetics include receptors of PTTH and ecdysis-related peptides (Palli & Cusson, 2007). An important factor in the development of new insecticidal compounds will be the identification of new target sites ranging from hormone receptors to hormone biosynthetic and degradative enzymes. Target site identification has been greatly helped by the complete sequencing of insect genomes, including species of economic importance, e.g. the honeybee (Apis mellifera) (Honeybee Genome Sequencing Consortium, 2006); and the silkworm (Xia et al., 2004 [Bombyx mori Biology Analysis Group]) as well as vectors of human diseases, e.g. the malaria mosquito (Anopheles gambiae) (Holt et al., 2002). Thus the full repertoire of peptidic hormones, and receptors for both peptidic and non-peptidic hormones, can eventually be known for a given sequenced insect genome by mining for genes with endocrine functions. This is exemplified by the discovery of 56 G-proteincoupled receptors (GPCRs) for neurohormones and biogenic amines in the honeybee genome (Hauser et al., 2006). Similarly, Drosophila genomics has made possible the identification of the suite of enzymes controlling ecdysone biosynthesis in this species (Gilbert & Warren, 2005). With large-scale DNA sequencing becoming more affordable, additional fully sequenced insect genomes should become available in the near future, including those of agricultural pests. This large-scale gene identification effort will feed into the strategies currently used in agrochemistry to synthesize and screen compounds against target sites (i.e. proteins). High-throughput instrumentation now allows the screening of 500 to 1000 compounds per day using in vitro assays on a given target (Allenza & Eldridge, 2007). In vitro assays, while allowing a higher throughput than whole-insect assays, are still trial-and-error operations. A target protein needs to be expressed in sufficient quantities and in the proper conformation to retain its in vivo functions in an in vitro context. Furthermore, some hormone receptors require two or more subunits to be functional (e.g. the ecdysone receptor), thus increasing the complexity and cost of some in vitro assays (Allenza &




Eldridge, 2007). To avoid these pitfalls, the discovery of novel hormone-based insecticidal compounds will require an intimate knowledge of the target at the molecular, cellular and wholeorganism levels. IPM specialists will likely witness the introduction of additional, endocrine-based control products in the foreseeable future. New approaches in the way endocrine disruption is delivered are also in development, for instance by using genetically modified plants or microorganisms that interfere with hormone action (Palli & Cusson, 2007). By exploiting the diversity and specificity of insect hormone systems, these new application methods should bring even more environmentally attractive control options for IPM.

References Addison, J. A. (1996). Safety testing of tebufenozide, a new molt-inducing insecticide, for effects on nontarget forest soil invertebrates. Ecotoxicology and Environmental Safety, 33, 55–61. Ahmad, M., Hollingworth, R. M. & Wise, J. C. (2002). Broad-spectrum insecticide resistance in obliquebanded leafroller Choristoneura rosaceana (Lepidoptera: Tortricidae) from Michigan. Pest Management Science, 58, 834–838. Allenza, P. & Eldridge, R. (2007). High-throughput screening and insect genomics for new insecticide leads. In Insecticide Design Using Advanced Technologies, eds. I. Ishaaya, R. Nauen & A. R. Horowitz, pp. 67–86. Berlin, Germany: Springer-Verlag. Altstein, M. (2004). Novel insect control agents based on neuropeptide antagonists: the PK/PBAN family as a case study. Journal of Molecular Neuroscience, 22, 147– 157. Anonymous (2002). Evaluation of the New Active Methoxyfenozide in the Product PRODIGY 240 SC Insecticide. Canberra, Australia: National Registration Authority for Agricultural and Veterinary Chemicals. Available at prsmeth.pdf/. Bombyx mori Biology Analysis Group, Xia, Q., Zhou, Z., Lu, C. et al. (2004). A draft sequence for the genome of the domesticated silkworm (Bombyx mori). Science, 306, 1937–1940. Brown, M. R., Raikhel, A. S. & Lea, A. O. (1985). Ultrastructure of midgut endocrine cells in the adult mosquito, Aedes aegypti. Tissue and Cell, 17, 709–721. ¨ ber die Isolierung Butenandt, A. & Karlson, P. (1954). U eines Metamorphosen-Hormons der Insekten in

kristallisierter Form. Zeitschrift fur Naturforschung, 9B, 389–391. Cadogan, B. L., Thompson, D. G., Retnakaran, A. et al. (1998). Deposition of aerially applied tebufenozide (RH-5992) on balsam fir (Abies balsamea) and its control of spruce budworm (Choristoneura fumiferana [Clem.]). Pesticide Science, 53, 80–90. Cadogan, B. L., Scharbach, R. D., Knowles, K. R. & Krause, R. E. (2005). Efficacy evaluation of a reduced dosage of tebufenozide applied aerially to control spruce budworm (Choristoneura fumiferana). Crop Protection, 24, 557–563. Castillo, M., Moya, P., Couillaud, F., Garcer´ a, M. D. & Martinez-Pardo, R. (1998). A heterocyclic oxime from a fungus with anti-juvenile hormone activity. Archives of Insect Biochemistry and Physiology, 37, 287– 294. Charrois, G. J. R., Mao, H. & Kaufman, W. R. (1996). Impact on salivary gland degeneration by putative ecdysteroid antagonists and agonists in the ixodid tick Amblyomma hebraeum. Pesticide Biochemistry and Physiology, 55, 140–149. Claeys, I., Simonet, G., Van Loy, T., De Loof, A. & Vanden Broeck, J. (2003). cDNA cloning and transcript distribution of two novel members of the neuroparsin family in the desert locust, Schistocerca gregaria. Insect Molecular Biology, 12, 473–481. Claeys, I., Breugelmans, B., Simonet, G. et al. (2006). Regulation of Schistocerca gregaria neuroparsin transcript levels by juvenile hormone and 20-hydroxyecdysone. Archives of Insect Biochemistry and Physiology, 62, 107– 115. Cusson, M. (2004). Juvenile hormone. In Encyclopedia of Entomology, ed. J. L. Capinera, pp. 1228–1230. Dordrecht, Netherlands: Kluwer. Cusson, M. & Palli, S. R. (2000). Can juvenile hormone research help rejuvenate integrated pest management? Canadian Entomologist, 132, 263–280. Cusson, M., B´eliveau, C., Sen, S. E. et al. (2006). Characterization and tissue-specific expression of two lepidopteran farnesyl diphosphate synthase homologs: implications for the biosynthesis of ethyl-substituted juvenile hormones. Proteins, 65, 742–758. Davey, M., Duve, H., Thorpe, A. & East, P. (2005). Helicostatins: brain-gut peptides of the moth, Helicoverpa armigera (Lepidoptera: Noctuidae). Archives of Insect Biochemistry and Physiology, 58, 1–16. Davis, N. T., Blackburn, M. B., Golubeva, E. G. & Hildebrand, J. G. (2003). Localization of myoinhibitory peptide immunoreactivity in Manduca sexta and Bombyx mori, with indications that the peptide has a role in molting and ecdysis. Journal of Experimental Biology, 206, 1449–1460.


Dhadialla, T. S., Carlson, G. R. & Le, D. P. (1998). New insecticides with ecdysteroidal and juvenile hormone activity. Annual Review of Entomology, 43, 545–569. Dhadialla, T. S., Retnakaran, A. & Smagghe, G. (2005). Insect growth- and development-disrupting insecticides. In Comprehensive Molecular Insect Science, vol. 6, eds. L. I. Gilbert, K. Iatrou & S. S. Gill, pp. 55–117. St Louis, MO: Elsevier. Dinan, L. (2001). Phytoecdysteroids: biological aspects. Phytochemistry, 57, 325–339. Dinan, L., Whiting, P., Girault, J. P. et al. (1997). Cucurbitacins are insect steroid hormone antagonists acting at the ecdysteroid receptor. Biochemical Journal, 327, 643–650. Doucet, D., Cusson, M. & Retnakaran, A. (2007a). Insect endocrinology and hormone-based pest control products in IPM. Table 1. In Radcliffe’s IPM World Textbook, eds. E. B. Radcliffe, W. D. Hutchison and R. E. Cancelado. Available at textbook/doucet1.htm/. Doucet, D., Cusson, M. & Retnakaran, A. (2007b). Insect endocrinology and hormone-based pest control products in IPM. Table 2. In Radcliffe’s IPM World Textbook, eds. E. B. Radcliffe, W. D. Hutchison and R. E. Cancelado. Available at doucet2.htm. Fussnecker, B. L., Smith, B. H. & Mustard, J. A. (2006). Octopamine and tyramine influence the behavioral profile of locomotor activity in the honeybee (Apis mellifera). Journal of Insect Physiology, 52, 1083– 1092. Gilbert, L. I. & Warren, J. T. (2005). A molecular genetic approach to the biosynthesis of the insect steroid molting hormone. Vitamins and Hormones, 73, 32–59. Gole, J. W. D. & Downer, R. G. H. (1979). Elevation of adenosine 3 , 5 -monophosphate by octopamine in fat body of the American cockroach Periplaneta americana L. Comparative Biochemistry and Physiology, 64C, 223–226. Hasegawa, K. (1957). The diapause hormone of the silkworm, Bombyx mori. Nature, 179, 1300–1301. Hauser, F., Cazzamali, G., Williamson, M., Blenau, W. & Grimmelikhuijzen, C. J. (2006). A review of neurohormone GPCRs present in the fruitfly Drosophila melanogaster and the honeybee Apis mellifera. Progress in Neurobiology, 80, 1–19. Helvig, C., Koener, J. F., Unnithan, G. C. & Feyereisen, R. (2004). CYP15A1, the cytochrome P450 that catalyzes epoxidation of methyl farnesoate to juvenile hormone III in cockroach corpora allata. Proceedings of the National Academy of Science of the USA, 101, 4024– 4029. Henderson J. (2005). Ernest Starling and “hormones”: an historical commentary. Journal of Endocrinology, 184, 5–10.

Holt, R. A., Subramanian, G. M., Halpern, A et al. (2002). The genome sequence of the malaria mosquito Anopheles gambiae. Science, 298, 129–149. Honeybee Genome Sequencing Consortium (2006). Insights into social insects from the genome of the honeybee Apis mellifera. Nature, 443, 931–949. Hsu, A. C.-T. (1991). 1,2-Diacyl-1-alkyl-hydrazines: a novel class of insect growth regulators. In Synthesis and Chemistry of Agrochemicals II, eds. D. R. Baker, J. G. Fenyes & W. K. Moberg, pp. 478–490. Washington, DC: American Chemical Society. Kethidi, D. R., Li, Y. & Palli, S. R. (2006). Protein kinase C phosphorylation blocks juvenile hormone action. Molecular and Cellular Endocrinology, 247, 127–134. ˇ nan, D., Cho, K. H. et al. (2006). Central pepKim, Y. J., Zitˇ tidergic ensembles associated with organization of an innate behavior. Proceedings of the National Academy of Science of the USA, 103, 14 211–14 216. King-Jones, K. & Thummel, C. S. (2005). Nuclear receptors: a perspective from Drosophila. Nature Reviews Genetics, 6, 311–323. Kogan, M. (1998). Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology, 43, 243–270. Kopec, S. (1917). Experiments on metamorphosis of insects. Bulletin of International Academy Cracov B, pp. 57–60. Kreutzweiser, D. P., Capell, S. S., Wainio-Keizer, K. L. & Eichenber, D.C. (1994). Toxicity of new molt-inducing insecticide (RH-5992) to aquatic macroinvertibrates. Ecotoxicology and Environmental Safety, 28, 14–24. Lagueux, M., Hoffmann, J. A., Goltzen´e, F. et al. (1984). Ecdysteroids in ovaries and embryos of Locusta migratoria. In Biosynthesis Metabolism and Mode of Action of Invertebrate Hormones, eds. J. A. Hoffmann & M. Porchet, pp 168–180. Heidelberg, Germany: Springer-Verlag. Medina, P., Smagghe, G., Budia, F., Tirry, L. & Vinuela, E. (2003). Toxicity and absorption of azadirachtin, diflubenzuron, pyriproxyfen, and tebufenozide after topical application in predatory larvae of Chrysoperla carnea (Neuroptera: Chrysopidae). Environmental Entomology, 32, 196–203. Minakuchi, C. & Riddiford, L. M. (2006). Insect juvenile hormone action as a potential target of pest management. Journal of Pesticide Science, 31, 77–84. Mommaerts, V., Sterk, G. & Smagghe, G. (2006). Bumblebees can be used in combination with juvenile hormone analogues and ecdysone agonists. Ecotoxicology, 15, 513–521. Monger, D. J., Lim, W. A., Kezdy, F. J. & Law, J. H. (1982). Compactin inhibits insect HMG-CoA reductase and juvenile hormone biosynthesis. Biochemical and Biophysical Research Communications, 105, 1374–1380.




Nakagawa, Y. (2005). Nonsteroidal ecdysone agonists. Vitamins and Hormones, 73, 131–173. Neves, C. A., Gitirana, L. B. & Serrao, J. E. (2003). Ultrastructure of the midgut endocrine cells in Melipona quadrifasciata anthidioides (Hymenoptera, Apidae). Brazilian Journal of Biology, 63, 683–690. Nijhout, H. F. (ed.) (1994). Insect Hormones. Princeton, NJ: Princeton University Press. Nordeen, S. K., Ogden, C. A., Taraseviciene, L. & Lieberman, B. A. (1998). Extreme position dependence of a canonical hormone response element. Molecular Endocrinology, 12, 891–898. Palli, S. R. & Cusson, M. (2007). Future insecticides targeting genes involved in the regulation of molting and metamorphosis. In Insecticides Design Using Advanced Technologies, eds. I. Ishaaya, R. Nauen & A. R. Horowitz, pp. 105–134. Berlin, Germany: SpringerVerlag. Palli, S. R., Tice, C. M., Margam, V. M. & Clark, A. M. (2005). Biochemical mode of action and differential activity of new ecdysone agonists against mosquitoes and moths. Archives of Insect Biochemistry and Physiology, 58, 234–242. Pratt, G. E., Kuwano, E., Farnsworth, D. E. & Feyereisen, R. (1990). Structure/activity studies on 1, 5-disubstituted imidazoles as inhibitors of juvenile hormone biosynthesis in aisolated corpora allata of the cockroach Diploptera punctata. Pesticide Biochemistry and Physiology, 38, 223–230. Predel, R., Eckert, M. & Holman, G. M. (1999). The unique neuropeptide pattern in abdominal perisympathetic organs of insects. Annals of the New York Academy of Science, 897, 282–290. Predel, R., Kellner R., Baggerman. G., Steinmetzer, T. & Schoofs, L. (2000). Identification of novel periviscerokinins from single neurohaemal release sites in insects: MS/MS fragmentation complemented by Edman degradation. European Journal of Biochemistry, 267, 3869–3873. Quistad, G. B., Cerf, D. C., Schooley, D. A. & Staal, G. B. (1981). Fluoromevalonate acts as an inhibitor of insect juvenile hormone biosynthesis. Nature, 289, 176– 177. Quistad, G. B., Cerf, D. C., Kramer, S. J., Bergot, B. J. & Schooley, D. A. (1985). Design of novel insect anti juvenile hormones: allylic alcohol derivatives. Journal of Agricultural Food Chemistry, 33, 47–50. Rachinsky, A., Mizoguchi, A., Srinivasan, A. & Ramaswamy, S. B. (2006). Allatotropin-like peptide in Heliothis virescens: tissue localization and quantification. Archives of Insect Biochemistry and Physiology, 62, 11–25.

Raikhel, A. S., Brown, M. B. & Belles, X. (2005). Hormonal control of reproductive processes. In Comprehensive Molecular Insect Science, vol. 3, eds. L. I. Gilbert, K. Iatrou & S. S. Gill, pp. 433–491. Amsterdam, Netherlands: Elsevier. Retnakaran, A., Smith, W. L., Tomkins, W. L. et al. (1997). Effect of RH-5992, a nonsteroidal ecdysone agonist, on the spruce budworm, Choristoneura fumiferana (Lepidoptera: Tortricidae): laboratory, greenhouse and ground spray trials. Canadian Entomologist, 129, 871–885. Retnakaran, A., Gelbic, I., Sundaram, M. et al. (2001). Mode of action of the ecdysone agonist tebufenozide (RH-5992), and an exclusion mechanism to explain resistance to it. Pest Management Science, 57, 951– 957. Retnakaran, A., Krell, P., Feng, Q. & Arif, B. (2003). Ecdysone agonists: mechanism and importance in controlling insect pests of agriculture and forestry. Archives of Insect Biochemistry and Physiology, 54, 187– 199. Riddiford, L. M., Hiruma, K., Zhou, X. & Nelson, C.A. (2003). Insights into the molecular basis of the hormonal control of molting and metamorphosis from Manduca sexta and Drosophila melanogaster. Insect Biochemistry and Molecular Biology, 33, 1327–1338. Robbins, W. E., Kaplanis, J. N., Thompson, M. J., Shortino, T. J. & Joyner, S. O. (1970). Ecdysones and synthetic analogs: molting hormone activity and inhibitive effects on insect growth, metamorphosis and reproduction. Steroids, 16, 105–125. S´ aenz-de-Cabez´ on Irigaray, F.-J., Marco, V., Zalom, F. G. & P´erez-Moreno, I. (2005). Effects of methoxyfenozide on Lobesia botrana Den & Schiff (Lepidoptera: Tortricidae) egg, larval and adult stages. Pest Management Science, 61, 1133–1137. Sakai, T., Satake, H., Minakata, H. & Takeda, M. (2004). Characterization of crustacean cardioactive peptide as a novel insect midgut factor: isolation, localization, and stimulation of alpha-amylase activity and gut contraction. Endocrinology, 145, 5671–5678. Sato, Y., Oguchi, M., Menjo, N. et al. (1993). Precursor polyprotein for multiple neuropeptides secreted from the suboesophageal ganglion of the silkworm Bombyx mori: characterization of the cDNA encoding the diapause hormone precursor and identification of additional peptides. Proceedings of the National Academy of Sciences of the USA, 90, 3251–3255. Sauphanor, B. & Bouvier J. C. (1995). Cross-resistance between benzoylureas and benzoylhydrazines in the coddling moth, Cydia pomonella L. Pesticide Science, 45, 369–375.


Sen, S. E., Trobaugh, C., B´eliveau, C., Richard, T. & Cusson, M. (2007). Cloning, expression and characterization of a dipteran farnesyl diphosphate synthase. Insect Biochemistry and Molecular Biology, 37, 1198–1206. Shinoda, T. & Itoyama, K. (2003). Juvenile hormone acid methyltransferase: a key regulatory enzyme for insect metamorphosis. Proceedings of the National Academy of Science of the USA, 100, 11 986–11 991. ˇ Sl´ ama K, Romaˇ nuk, M. & Sorm, F. (1974). Insect Hormones and Bioanalogues. New York: Springer-Verlag. Smagghe, G., Carton, B., Decombel, L. & Tirry, L. (2001). Significance of absorption, oxidation, and binding to toxicity of four ecdysone agonists in multi-resistant cotton leafworm. Archives of Insect Biochemistry and Physiology, 46, 127–139. Stanley, D. (2006). Prostaglandins and other eicosanoids in insects: biological significance. Annual Review of Entomology, 51, 25–44. Stay, B. & Tobe, S. (2007). The role of allatostatins in juvenile hormone synthesis in insects and crustaceans. Annual Review of Entomology, 52, 277–299. Sundaram, K. M. S., Nott, R. & Curry, J. (1996). Deposition, persistence and fate of tebufenozide (RH-5992) in some terrestrial and aquatic components of a boreal forest environment after aerial application of mimic.

Journal of Environmental Science, Health B, 31, 699– 750. Sundaram, M., Palli, S. R., Ishaaya, I., Krell, P. J. & Retnakaran, A. (1998). Toxicity of ecdysone agonists correlates with the induction of CHR3. Pesticide Biochemistry and Physiology, 62, 201–208. Truman, J. W. & Copenhaver, P. F. (1989). The larval eclosion hormone neurones in Manduca sexta: identification of the brain-proctodeal neurosecretory system. Journal of Experimental Biology, 147, 457–470. Wearing, C. H. (1998). Cross-resistance between azinophosmethyl and tebufenozide in the greenheaded leaf roller, Planotortrix octo. Pesticide Science, 54, 203–211. Williams, C. M. (1956). The juvenile hormone of insects. Nature, 178, 212–213. Williams, C. M. (1967). Third-generation pesticides. Scientific American, 217, 13–17. ˇ nan, D. & Adams, M. E. (2005). Neuroendocrine regulaZitˇ tion of insect ecdysis. In Comprehensive Molecular Insect Science, vol. 3, eds. L. I. Gilbert, K. Iatrou & S. S. Gill, pp. 1–60. Amsterdam, Netherlands: Elsevier. ˇ nan, D., Zitˇ ˇ nanov´ Zitˇ a, I., Spalovsk´ a, I. et al. (2003). Conservation of ecdysis-triggering hormone signalling in insects. Journal of Experimental Biology, 206, 1275– 1289.


Chapter 23

Eradication: strategies and tactics Michelle L. Walters, Ron Sequeira, Robert Staten, Osama El-Lissy and Nathan Moses-Gonzales There are four main terms with similar yet unique definitions to consider when developing a method of pest management. Eradication is the application of phytosanitary measures to eliminate a pest from an area or geographic region (Food and Agriculture Organization, 2005). Suppression involves maintaining an insect population at or below the economic injury level (Pfadt, 1972; Hendrichs et al., 2002). Containment is the application of phytosanitary measures in and around an infested area to prevent spread of a pest (FAO, 2005). Prevention is the application of phytosanitary measures in and/or around a pest free area to avoid the introduction of a pest. Of the four, only prevention involves a preemptive strategy to keep the pest at bay rather than managing it upon arrival (Hendrichs et al., 2002, 2005; Food and Agriculture Organization, 2005). The other three terms, eradication, suppression and containment, are designed to manage a pest after its initial infestation. These terms are not mutually exclusive; rather, they provide different strategies and tactics designed to custom fit an areawide IPM (AWIPM) program. The section of this chapter on pink bollworm (Pectinophora gossypiella) provides a brief review of San Joaquin Valley, California in regards to prevention, as well as the ongoing efforts to eradicate this invasive lepidopteran insect from the cotton growing regions of the southwest USA and northern Mexico. The section on Mediter-

ranean fruit fly (Ceratitis capitata) of this chapter discusses in greater detail the economic viability and opportunity cost of eradication versus suppression in the international market. The final section of this chapter discusses the history, efficiency and effect of the boll weevil (Anthonomus grandis) eradication program. “We have,” according to Klassen and Curtis (2005), “entered an era of an unprecedented level of travel by exotic invasive organisms.” As a result, we have also entered an era of unprecedented research and development of means and methods designed to reduce and eradicate these exotic invasive organisms. This chapter explores the principles of insect eradication, as well as the means and methods necessary to achieve this daunting task. For our purposes, the term “eradication” is defined as “A type of regulatory-control program in which a target pest is eliminated from a geographical region (Gordh & Headrick, 2001)” versus the one traditionally used for infectious disease, “extinction of the pathogen in humans and/or the environment” (Arita et al., 2004). The reader is directed to Klassen (1989) Eradication of Introduced Arthropod Pests: Theory and Historical Practice for an excellent review of the practice in the USA. There is no standard outline or guide to running an eradication program. Instead, many programs rely on organic support structures rather than rigid systems.

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 


Motivations for eradication campaigns vary with the situation and pest. Klassen (1989) states that the presence of new pests may bother people directly (bite or sting), vector diseases to humans, livestock or crops, cause extensive direct damage to crops or the environment, cause trade embargos or significantly increase the use of pesticides and consequently damage the environment. To date, exotic pests (insects, mammals, mollusks, weeds and diseases) are responsible for an estimated one third of agricultural losses in the USA. Natural ecosystems may also be adversely affected by invasive exotics. Dutch elm disease (Ophiostoma novo-ulmi), chestnut blight (Cryphonectria parasitica), pink bollworm, red imported fire ant (Solenopsis invicta) and Kudzu vine (Pueraria lobata) are but a few examples. People are left with the desire to eradicate the pest, whether introduced or not, as in the case of expanded ranges of boll weevil and screwworm (Cochliomyia hominivorax). The practice of eradication is controversial, as well it should be, due to the uncompromising nature of the practice, the uncertainty of success and the extreme measures needed to achieve it. Moral issues around the eradication of a native species and unintended consequences from eradication efforts including damage to the environment, loss of non-target species including pollinators and rare, native insects (Knipling, 1978; Rabb, 1978) are just a few of the possible negative impacts of eradication programs.

23.1 Is eradication an option? Klassen concluded that the justification for eradication must be based on the “anticipated economic, ecological, and sociological consequences.” In analyzing program efficiency, one must heed the warnings communicated by Myers et al. (1998) and avoid known biases and shortcomings commonly associated with economic analyses. Myers et al. warn that the benefits of eradication programs are “almost always overestimated” because of lack of scientific data to separate impacts from an eradication program from other, unrelated factors, potentially biased decision processes due to strong stakeholders in industry, or potentially biased evaluations due to

primary focus on producers. At the same time, they warned that “in contrast to benefits, biases in procedures often underestimate the costs of proposed eradication programs.” Examples of such costs include the escalating expense of eradicating the last individuals, unanticipated impacts on other aspects of society, the need for continuous monitoring of populations, risks of potential reintroductions, public relations, potential lawsuits, costs of human error and risks to human health. With the above caveats, economic analyses may provide information about efficiencies, not about program efficacy. That is an important distinction.

23.1.1 Smallpox prototype While the focus of this chapter will be on the principles of eradicating an insect from a geographical region, smallpox (Variola vera, an infectious disease) provides a prototype of successful eradication. In 1980, smallpox became the first, and to date, only successfully eradicated human disease. The effort took 96 years and was aided by numerous characteristics of the disease. Humans were the sole reservoir, thus immunization of humans stopped transmission of the disease. Patients did not shed the virus after recovery. There were no subclinical infections, that is, infected people displayed symptoms, and disease symptoms were easily and accurately identified, making surveillance effective. The treatment, a vaccine, was nearly 100% effective. In addition, the smallpox eradication program was strongly led by the United Nations World Health Organization and carried out by cooperative member states with funding and logistics, and in a time of relative political stability (Arita et al., 2004). The smallpox prototype serves as a launching pad for the main factors that ensure successful insect eradication programs. The prerequisites for developing an eradication strategy requires an understanding of life cycle, climatic needs, geographic distribution, mating habits, preadaptation to its new environment, degree of genetic plasticity, number of generations per year, reproductive capacity, ability to compete for niches, natural enemies and host species. The implementation of programs typically requires the examination, evaluation and/or development of a wide variety of tactics, including




agronomic practices, and the availability of powerful methods to suppress the pest, including resistant hosts (through selective breeding or genetic modification), sterile insect technique (SIT), mating disruption, pesticide treatments, insect growth regulators, biological control, host elimination, and cultural controls including hostfree periods and crop diversity. Early detection of introduced species, indicating an effective means of survey immediately followed by a decisive plan of action (methods) is beneficial to control losses and minimize costs. In addition to those listed by Klassen (1989), successful eradication benefits from a well-defined and geographically delimited host range. This requires the implementation of new technologies (GPS, GIS, etc.) in tandem with conventional survey techniques (field surveying, direct observation, etc.). Geospatially referenced data, stored in well-designed databases, facilitates documentation and tracking supports decision making surrounding the eradication effort. Also required are survey techniques that gather reliable estimation of pest presence and population size. The support of large and robust infrastructures is necessary to fund and carry out eradication programs. Broad economic and political support, i.e. areawide effort, along with good communication, record keeping and evaluation are necessary to ensure that program needs are addressed quickly, as in the case of boll weevil in the cotton belt. Klassen (1989) attributes the establishment of eradication as a strategy to C. H. Fernald of University of Massachusetts who was the architect behind the widespread use of the arsenical pesticide Paris Green and other tactics to successfully contain and eradicate the gypsy moth (Lymantria dispar) from 1890 to 1901. Local containment was achieved; however, eradication failed due to poor public support and lack of foresight from the federal government. To date, gypsy moth has not been eradicated. Public perception and education is essential to ensure continuing public support of an eradication program. Education includes outreach programs, publications and the use of mass media (e.g. the internet, television, radio, newspapers, etc.). The public should be informed about the economic costs/benefits of the program, as well as the health and quality of life they may enjoy post-

eradication. If eradication is merited and public education is actively pursued, then cooperation by the public should follow suit. “Adequacy of legal authority” is vital not only to conduct the eradication program, but also for effective regulation, inspection and quarantine programs that are geared to prevent reinfestation. In order to initiate an eradication program successfully, a stakeholder-supported, areawide organization is required. The organization must not be inhibited by geopolitical borders and benefits by boundaries with strong means of enforcement, such as geographic properties that isolate it from rapid reintroduction of the pest. Furthermore, “Cohesiveness of stakeholders in the private sector and effective stakeholder leadership,” as well as “strong support of political leaders” is vital to a successful eradication program (Klassen, 1989). Finally, realizing that eradication is not always a permanent venture is paramount for an eradication program. A successful program requires longterm maintenance and monitoring systems. Without these systems, the fruits of eradication may be wiped out. Achieving eradication, according to Klassen (1989), requires a zero count of insects for at least ten generations. In order to ensure and maintain a zero insect count, continual surveying, as well as a contingency plan that would successfully eliminate any possible reintroduction of the insect is required.

23.2 Case studies 23.2.1 Pink bollworm The worldwide spread of pink bollworm from its presumed origin in India is summarized by Ingram (1994). The history of pink bollworm in the USA is summarized by Noble (1969). In brief, pink bollworm arrived in the western hemisphere in infested seed shipped from Egypt to Mexico in 1911 and is currently the key pest of cotton across the Southwest of the USA and northern Mexico. Lepidoptera are among the most destructive insect pests in the world. According to Klassen & Curtis (2005) “lepidopteran larvae cause immense damage to food and forage crops, forests, and stored products.” Pink bollworm feeds almost exclusively on cotton (Gossypium spp.) and can cause devastating economic loss by dramatically


reducing the yield and quality of cotton lint (Pfadt, 1978). The larvae bore into the developing cotton fruit, where they feed on the cotton lint and seeds. The pink bollworm is difficult to control with insecticides because the egg is typically protected under the calyx of the boll and it spends the destructive larval phase inside the cotton boll where it is also well protected. Cultural controls, such as a short growing season, enforced plowdown and a host-free period, have successfully decreased populations (Chu et al., 1992) and are widely used but are insufficient to prohibit economic loss. In years past, pink bollworm has cost growers in counties of Texas as much as $US 52 million in a single season (Pfadt, 1972). Currently, infestations by pink bollworm costs USA cotton producers over $US 32 million each year in control costs and yield losses (National Cotton Council, 2007). In addition to production losses, there are quarantine implications that limit cotton export markets. Efforts to control damage by pink bollworm began early on. According to James Rudig (Program Manager, California Department of Food and Agriculture), “the Pink Bollworm Program in the San Joaquin Valley is probably the most successful and longest running yet least known areawide Integrated Pest Control (IPC) program in the world.” The program has been in continual operation since 1967. Currently the program uses mapping of cotton fields, trapping, population modeling, cultural practices, sterile insect release and occasional use of mating disruption to prevent the establishment of pink bollworm. The program is supported by bale assessments with some federal government support (California Department of Food and Agriculture, 2007). A six-year (1989–95) areawide program to control pink bollworm in Parker, Arizona demonstrated how careful mapping, trapping and prompt application of pheromone mating disruption could drive a heavy pink bollworm infestation to near zero in that valley (Antilla et al., 1996). Similarly, from 1994 to 2000, a research trial was conducted to determine if pink bollworm could be controlled using the Sterile Insect Technique (SIT), in the heavily infested Imperial Valley of California. Early successes (and a few failures) showed promise and the program was expanded to include the also heavily infested Palo

Verde Valley on the California–Arizona border. SIT is advantageous to pink bollworm eradication because the insect is released during the benign, adult phase of its life cycle when it feeds innocuously on nectar. Naturally, mating occurs during the adult portion of the pink bollworm’s life cycle. Thus, releasing sterilized adult pink bollworm does not cause further damage to the cotton and yet reduces the amount of progeny. During the course of the SIT trial, Bacillus thuringiensis (Bt) cotton came into widespread use and was incorporated in the experiment. The trial had proven its point by 1999 – SIT alone or SIT with Bt cotton could suppress pink bollworm populations; however, the eradication plan did not pass in Arizona and the trial was ended (Pierce et al., 1995; Staten et al., 1999; Walters et al., 2000). As in Parker, in the year following the end of each trial, pink bollworm populations rebounded (Walters et al., 2000). After its faltering start in the West, by 2002 pink bollworm eradication was active in Texas and New Mexico, following quickly after the Boll Weevil Eradication Program. Currently, pink bollworm is under eradication in the USA, including the states of Texas, New Mexico, Arizona and California as well as in the northern Mexico state of Chihuahua. Although pink bollworm eradication was achieved at various times and in specific locations, the development of new technologies, such as mating disruption, transgenic crops and SIT, has once again made pink bollworm eradication economically feasible. The Pink Bollworm Eradication Program requires multi-year support at the grower and federal level, as well as willingness to remain committed to the program (Fig. 23.1). The Pink Bollworm Eradication Program and others like it are funded by a congressional line item. While the US Federal Government provides $US 4–6 million a year for sterile insect rearing and release in the eradication program, the governmental support of the program costs is relatively small compared to the growers’ 80% share in the funding (National Cotton Council, 2007). The government also provides scientific support and a framework for the program. Frisvold (2006) contends that, given current economic trends, the current price of seed, insecticide and technological costs, the Pink Bollworm Eradication Program will become




Fig. 23.1 Incremental phases of the APHIS Pink Bollworm Eradication Program.

cost efficient in five years, if non-Bt cotton is purchased, and no technology fees are assessed, or in six years if Bt cotton is actively purchased and technology fees are assessed. Early infestations of pink bollworm in the USA underwent cycles of eradication and reintroduction as various counties came under and went out of quarantine regulations. Past government initiatives, such as the Pink Bollworm Act of 1918 did not have the legal backing to enforce cultural and sanitary controls. As a result, cotton growers refused to follow the government programs and law enforcement was unable, and sometimes unwilling, to enforce laws. While many cotton growers wanted pink bollworm out of their fields, infestations continued to plague growers because they were unwilling to follow the Pink Bollworm Act of 1918 (Geong, 2000). Grower ambivalence was a result of a two-prong failure, namely the lack

of adequate educational programs established to explain why eradication was vital to the cotton industry and the lack of governmental foresight that ensured adequate legal authority to successfully carry out eradication. Thus, government foresight and adequate policies to ensure perennially funded and implemented programs are paramount to the success of the present pink bollworm eradication effort or any other eradication program. Finally, the crux of any eradication program undertaken on cultivated crops is the predominant role of the grower over governmental authority. This ensures that the grower takes responsibility and provides the critical agricultural input that takes the program out of the laboratory and into the field.

23.2.2 Mediterranean fruit fly The Mediterranean fruit fly, commonly referred to as “medfly,” has been described as the world’s most threatening agricultural pest, attacking over 200 different fruits, vegetables and nuts (Thomas et al., 2005). Medfly females deposit their eggs in


the epi- or mesocarp region of ripening host fruits. The eggs are laid in clutches of one to ten eggs (up to 800 during its life), larvae pass through three instars feeding on the fruit; they leave the fruit to pupate in the soil. After emergence and before becoming sexually active, adults feed on carbohydrates and water to survive and on protein sources to allow for gonad maturation (Christenson & Foote, 1960; Weems, 1981). The time required to complete a life cycle under summer weather conditions (e.g. in Florida) is 21–30 days. Females die soon after ceasing to oviposit. Development of pupae is variable and an induced quiescence may help this pest survive through unfavorable conditions. Adults can fly short distances, but winds may carry them 2 km or more. According to reports, 50% of the flies that emerge die during the first two months of life; however, adults may survive up to a year or more under favorable conditions. When host fruit is continuously available and weather conditions favorable for many months, successive generations can be large and of mixed age distribution. Lack of fruit for three or four months reduces the population dramatically (Back & Pemberton, 1915, 1918; Weems, 1981). The largely internal life cycle of the feeding immature stages is key to understanding the challenges associated with medfly control. From their likely origin in sub-Saharan Africa, medflies spread rapidly around the world (Davies et al., 1999; Thomas et al., 2005) largely by human trade in fresh fruits and vegetables. Its frequent incursions worldwide have demonstrated the medfly’s destructive capacity and extended host range. In addition to its direct impact in reducing yields, medfly is a quarantine pest for the USA and for many important trading partners. The presence of medfly in a given area implies that fruit products moving out of that area (exports and even domestic movement if the fly population is contained and restricted) need to be treated with expensive phytosanitary treatments such as hot water immersion, fumigation, irradiation cold treatment. Often, the only cost-effective measure is to cease exports to countries that have quarantines against medfly. Medflies reached the Americas by 1900. Medflies were first reported from Hawaii in 1907 (Harris, 1989), from Florida in 1929, from California in 1975 (Metcalf, 1995; Thomas et al.,

2005) and from Texas in 1966. Whereas considered established in Hawaii, medfly has been officially declared eradicated from the continental USA despite periodic outbreaks that have occurred in Florida and California since 1975. Scientists (e.g. Carey, 1991) suggested that California may have an undetectable resident population. Whether eradicated, or established but undetectable, California and Florida have adopted permanent SIT programs to ferret out remaining flies and prevent measurable reinfestation. Medfly and economic viability: international benefits of eradication Each new country that is invaded by a spreading disease or pest represents a new node for the invasive organism and adds strength to the invasion process, complicating global management efforts. An analogy is made to the internet where each node (server) adds redundancy and stability to a highly interconnected system. The fact that trade and human traffic have become so widespread strengthens the analogy between the internet and epidemiology. An infestation in a given country has direct implications to the likelihood of spread to all other countries − increasing probability of spread. In the following equation, Global Welfare (GW) is the value to all nations engaged in open trading systems that accrue from the reduced probability of pest introduction that result from eradication of a pest in a given nation.  G Wi =


Export value for host commodities  Global export value i, j

where i = country index j = commodity index. The fruit fly GWi value is the proportion of total commerce in fruit fly susceptible commodities that is consumed by a given country (i). The numerator is the theoretical maximum for investment in global eradication efforts. In practice the maximum investment in eradication is approximated by the actual losses due to fruit fly. A constant global value is used here to minimize the variation across commodities and conditions, such that:




Maximum eradication investmenti 



Export value for host commodities  Global export value i, j

∗ Global lossesl

where l = year index. Despite the generalizations (e.g. we do not include domestic pest management or eradication costs), those expressions can be used to explore the value that eradication in a third country has for one’s own country. It can also be used to determine (in a global economy) what share of the total eradication costs in a country with limited sanitary and phytosanitary infrastructure should be borne by countries that benefit from exports in that commodity and are susceptible to that pest (fruit flies in this example). This is important because when national plant protection organizations conduct cost–benefit studies, the benefits considered are only those that accrue to an individual country or even an individual state, such as Hawaii (Mumford, 2005). In the GW context, however, the benefits of the absence of fruit flies in Hawaii are enjoyed mostly by mainland America. The concept of GW is applied to fruit flies here but is of general applicability. International benefits of regional suppression The statutory discretion of US Department of Agriculture Animal Plant Health Inspection Service (APHIS) was established in part with a view to increase and sustain exports as well as to ensure that when trade occurs, phytosanitary risks are minimized. The establishment of officially recognized, managed “low prevalence areas” could be beneficial in that there is clearly great potential to integrate with systems approaches in support of expanded export markets. The use of areawide suppression approaches is compatible with both of those guidelines (Food and Agriculture Organization, 2006). The analysis of effectiveness and efficiency indicators confirms technological viability of eradication and areawide suppression (Dyck et al., 2005), albeit with eradication having greater uncertainties and technological limitations than areawide suppression. The analyses of efficiency

show the areawide suppression options routinely have significantly lower initial costs, but higher long-term maintenance costs than the eradication option. However, the assumption that eradication is “forever” runs counter to recent history with more than 15 outbreaks of tephritid fruit flies in the past two decades in the USA. Eradication always shows the highest benefits over the long term (compared to areawide management or the status quo), if one assumes that no new introductions occur. If eradication is not maintained over the long term, then costs can overrun benefits depending on the frequency of new introductions. Consequently, suggestions have been forwarded that the fruit fly control program in Hawaii should shift from eradication and toward areawide management (Mitchell & Saul, 1990). There is ample evidence that areawide programs for fruit fly management are successful in reducing populations, allowing increased yields, and in some cases, permitting the export of fruit under specific quality controls, monitoring and implementation of appropriate international standards such as cited above.

23.2.3 Boll weevil The boll weevil, a native of Mexico and Central America, was first introduced into the USA near Brownsville, Texas, in about 1892 (Hunter, 1905; Hunter & Hinds, 1905). By 1922, the weevil had spread into cotton-growing areas of the USA from the eastern two-thirds of Texas and Oklahoma to the Atlantic Ocean. Northern and western portions of Texas were colonized by the boll weevil between 1953 and 1966 (Newsom & Brazzel, 1968). The history of boll weevil and its effects on the cotton industry as well as efforts to manage or eradicate the boll weevil have been reviewed by Klassen (1989), El-Lissy et al. (1996), Haney et al. (1996) and Dickerson et al. (2001). The Boll Weevil Eradication Program in the USA began in 1893, or one year after the boll weevil was first observed in Brownsville. The boll weevil was immediately recognized as an invasive, though not exotic pest species, which was expanding its range northward from Mexico into Texas and the southern states. Fast action was taken and various methods were implemented; however, these early efforts failed to stop the spread of


boll weevil. While the boll weevil was an extremely devastating pest economically, it contributed to the westward spread of the cotton industry and to the twentieth century shift in the southeast from a cotton monoculture towards a wider variety of crops. The presence of boll weevil also contributed greatly to the development of entomology in the southern USA (Hardee & Harris, 2003). A monument was created in Enterprise, Alabama to commemorate the role of the boll weevil in forcing crop diversity in the region. The boll weevil’s main host is cotton. Boll weevils survive winter as diapausing adults (Hardee & Harris, 2003). Nearly 50% of female boll weevils can store sperm, thus eliminating the necessity to find a mate before laying eggs after re-emergence (Beckham, 1962) so that even harsh winters, where populations may be severely depleted tend not to eliminate the population. These characteristics contribute to the invasive nature of the boll weevil. Early suppression methods included the aerial application of arsenic dusts and chlorinated hydrocarbons; however, by the 1950s boll weevil developed resistance to these tactics (Hardee & Harris, 2003). In view of the economic and environmental problems posed by the boll weevil and its control, and in recognition of the technical advances developed during 80-plus years of research, most notably by E. F. Knipling, J. R. Brazzel and T. B. Davich, a cooperative boll weevil eradication experiment was initiated in 1971 in southern Mississippi and parts of Louisiana and Alabama (Parencia, 1978; Perkins, 1980). “The ultimate goal” of this project “was to discover or develop a tactic or combination of tactics that would provide a long-term solution to the boll weevil problem in cotton production” (Hardee & Harris, 2003). This experiment used an areawide IPM approach including chemical control, release of sterile male weevils, mass trapping and cultural controls. Based on this experiment, the National Cotton Council of America concluded that it was technically and operationally feasible to eliminate the boll weevil from the USA. By 2007, the boll weevil had been eradicated from nearly 5.3 million hectares of cotton in: Virginia, North Carolina, South Carolina, Georgia, Florida, Alabama, Kansas, California and Arizona,

and portions of Tennessee, Mississippi, Missouri, Arkansas, Louisiana, Oklahoma, Texas and New Mexico, as well as from the neighboring regions of the Mexicali Valley, Sonoita and Caborca in Mexico (Fig. 23.2). The program is currently operating in the remaining 1.42 million hectares of cotton in Tennessee, Mississippi, Missouri, Arkansas, Louisiana, Oklahoma, Texas and New Mexico. As of 2007, 100% of the USA Cotton Belt is involved in boll weevil eradication, with nearly 80% having completed eradication and the remaining 20% nearing eradication. Nationwide eradication in the USA is expected in 2008. The operational success of the current eradication program hinges on three interdependent components: mapping, detection and decisions resulting in well-timed application of control methods. Mapping is one of the first phases of operation. Mapping identifies the exact location of each cotton field and defines the surrounding environment. Additionally, each field is identified with a unique number to provide for accurate data management. Chemical control consists of a single aerial application of malathion (ultralow-volume, ULV) beginning at the pinhead-square growth stage to fields that had reached the treatment criteria (action threshold). By preemptively targeting pinhead-squares before sustainability is achieved, adult emergence numbers are reduced (Hardee & Harris, 2003). The 2004 season-long action threshold for treatment was a trap catch of 1–2 adult boll weevils per field (16 ha or less) in all active zones. Chemical treatment is the principal method of suppressing the boll weevil population. All eradication zones use the boll weevil pheromone trap as the primary tool of detection (Cross, 1973; El-Lissy et al., 1996). Although the primary function of the trap is detection, an indirect benefit of trapping, especially in low weevil populations, is that it removes a percentage of the population (Lloyd et al., 1972). The Boll Weevil Eradication Program provides an amalgamation of various techniques including automated trap data collection and a database designed to provide decision support. Time-frames for uniform cotton planting and harvesting, as organized by growers, local agricultural extension services, and in some cases state regulatory agencies, are key components of




Fig. 23.2 Progress of the APHIS Boll Weevil Eradication Program.

cultural control in providing the necessary hostfree period. In some states such as Arkansas and Texas, growers were offered a rebate to destroy crop residues as soon as possible after harvest in an effort to reduce overwintering populations and insecticide treatments. During the 2005 diapause phase, the Rio Grande Valley and the Northern Blacklands of Texas underwent weekly aerial applications with malathion ULV. These treatments began on 15 June and 15 July, respectively, and continued until cotton fields were defoliated and harvested. Trapping in post-eradication zones provides early warning of boll weevil reintroduction, from natural migration or artificial movement. Early detection allows an immediate response in containing and eradicating the reintroduced population before it reestablishes. Post-eradication trap-

ping will continue until nationwide eradication is complete, at which time a reduced trapping density will be put in place.

23.3 Conclusions Success of an eradication program depends on an overwhelming desire to eradicate an invasive insect pest. Effective programs such as the Boll Weevil Eradication Program experience perennial public and governmental support and funding and were promoted by educational outreach and stakeholder involvement. Well-documented research and dedicated professionals in both the scientific and farming communities help lead to a clear, well-defined program with a solid foundation of knowledge regarding to an insect pest’s biology, life cycle, climatic needs, mating behaviors, geographic distribution, etc. Finally, the confluence of desire, availability of suitable methods, funding, identification of weak links in the armor


of the pest, tremendous hard work and an ounce of luck are the keys to a successful eradication program.

References Antilla L., Whitlow, M., Staten, R. T., El-Lissy, O. & Myers, F. (1996). An integrated approach to areawide pink bollworm management in Arizona. In Proceedings of the Beltwide Cotton Conference, 2, pp. 1083–1085. Memphis, TN: National Cotton Council. Arita, I., Wickett J. & Nakane, M. (2004). Eradication of infectious diseases: its concept, then and now. Japanese Journal of Infectious Diseases, 57, 1–6. Back, E. A. & Pemberton, C. E. (1915). Life history of the Mediterranean fruit fly from the standpoint of parasite introduction. Journal of Agricultural Research, 3: 363–374. Back, E. A. & Pemberton, C. E. (1918). The Mediterranean Fruit Fly, US Department of Agriculture Bulletin No. 640. Washington, DC: US Government Printing Office. Beckham, C. M. (1962). Seasonal Studies of Diapause in the Boll Weevil in Georgia, Mimeograph Series N.S. No. 161. Athens, GA: University of Georgia, Georgia Agricultural Experiment Station. California Department of Food and Agriculture (2007). Pink Bollworm: Program Details. Sacramento, CA: California Government. Available at ipc/pinkbollworm/pbw_hp.htm. Carey, J. R. (1991). Establishment of the Mediterranean fruit fly in California. Science, 253, 1369–1373. Christenson, L. D. & Foote, R. H. (1960). Biology of fruit flies. Annual Review of Entomology, 5, 171–192. Chu, C. C., Weddle, R. C., Staten, R. T. et al. (1992). Pink bollworm: populations two years following initiation of a short-season cotton system in the Imperial Valley, CA. In Proceedings: Beltwide Cotton Production and Research Conference, pp. 804–806. Memphis, TN: National Cotton Council. Cross, W. H. (1973). Biology, control and eradication of the boll weevil. Annual Review of Entomology, 18, 17– 46. Davies, N., Villablanca, F. X. & Roderick, G. K. (1999). Bioinvasions of the Medfly Ceratitis capitata: source estimation using DNA sequences at multiple intron loci. Genetics, 153, 351–360. Dickerson, W. A., Brashear, A. L. Brumley, J. T. et al. (2001). Boll Weevil Eradication in the United States through 1999. Memphis, TN: Cotton Foundation. Dyck, V., Hendrichs, J. & Robinson, J. (eds.) (2005). Sterile Insect Technique. New York: Springer-Verlag.

El-Lissy, O., Myers, F., Frisbie, R. et al. (1996). Boll weevil eradication status in Texas. In Proceedings of the Beltwide Cotton Production and Research Conference, pp. 831–839. Memphis, TN: National Cotton Council. Food and Agriculture Organization (2005). International Standards for Phytosanitary Measures: Glossary of Phytosanitary Terms (updated 2007), ISPM No. 5. Produced by the Secretariat of the International Plant Protection Convention. Rome, Italy: Food and Agriculture Organization of the United Nations. Food and Agriculture Organization (2006). International Standards for Phytosanitary Measures (ISPM), 2005 edn. Produced by the Secretariat of the International Plant Protection Convention. Rome, Italy: Food and Agriculture organization of the United Nations. Available at Frisvold, G. (2006). Economics of Pink Bollworm Eradication. Memphis, TN: Cotton Incorporated. Geong, H.-G. (2000). The pink bollworm campaign in the South: agricultural quarantines and the role of the public in insect control, 1915–1930. Agricultural History, 74, 309–321. Gordh, G. & Headrick, D. H. (eds.) (2001). A Dictionary of Entomology. Wallingford, UK: CABI Publishing. Haney, P. B., Lewis, W. J. & Lambert, W. R. (1996). Cotton Production and the Boll Weevil in Georgia: History, Cost of Control, and Benefits of Eradication, Research Bulletin No. 428. Athens, GA: University of Georgia, Georgia Agricultural Experiment Station. Hardee, D. D. & Harris, F. A. (2003). Eradicating the boll weevil. American Entomologist, 49, 82–97. Harris, E. J. (1989). Pest status in Hawaiian islands and North Africa. In World Crop Pests: Fruit Flies, Their Biology, Natural Enemies and Control, vol. 3A, eds. A. Robinson & G. Hooper, pp. 73–80. Amsterdam, Netherlands: Elsevier. Hendrichs, J., Robinson, A. S., Cayol, J. P. & Enkerlin, W. (2002). Medfly areawide sterile insect technique programmes for prevention, suppression or eradication: the importance of mating behavior studies. Florida Entomologist, 85, 1–13. Hendrichs, J., Vreysen, M. J. B., Enkerlin, W. R. & Cayol, J. P. (2005). Strategic options in the use of the sterile insects for area-wide integrated pest management. In Sterile Insect Technique, eds. V. Dyck, J. Hendrichs & J. Robinson, pp. New York: Springer-Verlag. Hunter, W. D. (1905). The Control of the Boll Weevil, Including Results of Recent Investigations, US Department of Agriculture Farmers’ Bulletin No. 216. Washington, DC: US Government Printing Office. Hunter, W. D. & Hinds, W. E. (1905). The Mexican Cotton Boll Weevil, USDA Agricultural Handbook




No. 512. Washington, DC: US Government Printing Office. Ingram, W. R. (1994). Pectinophora (Lepidoptera: Gelechiidae). In Insect Pests of Cotton, eds. G. A. Mathews & J. P. Tunstall, pp. 107–148. Wallingford, UK: CAB International. Klassen, W. (1989). Eradication of Introduced Arthropod Pests: Theory and Historical Practice, Miscellaneous Publication No. 73. Lanham, MD: Entomological Society of America. Klassen, W. (2005). Area-wide integrated pest management and the sterile insect technique. In Sterile Insect Technique, eds. V. A. Dyck, A. J. Hendrichs & A. S. Robinson, pp. 39–68. New York: Springer-Verlag. Klassen, W. & Curtis, C. F. (2005). History of the Sterile Insect Technique. In Sterile Insect Technique, eds. V. Dyck, J. Hendrichs & J. Robinson, pp. 3–36. New York: Springer-Verlag. Knipling, E. F. (1978). Strategic and tactical use of movement information in pest management. In Conference on Radar, Insect Population Ecology and Pest Management, NASA Conference Publication No. 2070, pp. 41–57. Lloyd, E. P., Merkl, M. E., Tingle, F. C. et al. (1972). Evaluation of male-baited traps for control for boll weevils following a reproduction-diapause program in Monroe County, Mississippi. Journal of Economic Entomology, 65, 552–555. Metcalf, R. L. (1995). Biography of the Medfly. In The Medfly in California: Defining Critical Research, eds. J. G. Morse, R. L. Metcalf, J. R. Carey & R. V. Dowell, pp. 43–48. Riverside, CA: University of California, Center for Exotic Pest Research. Mitchell, W. C. & Saul, S. H. (1990). Current control methods for Mediterranean fruit fly, Ceratitis capitata, and their application in the USA. Review of Agricultural Entomology, 78, 923–940. Mumford, J. D. (2005). Application of benefit/cost analysis to insect pest control using the sterile insect technique. In Sterile Insect Technique, eds. V. Dyck, J. Hendrichs & J. Robinson, pp. 481–498. New York: Springer-Verlag. Myers, J. H., Savoie, A. & van Randen, E. (1998). Eradication and pest management. Annual Review of Entomology, 43, 471—491. National Cotton Council (2007). Pink Bollworm Eradication: Proposal and Current Status. Available at www. Newsom, L. D. & Brazzel, J. R. (1968). Pests and their control. In Advances in Production and Utilization of Quality Cotton: Principles and Practices, eds. F. C. Eliot,M. Hoover

& W. K. Porter Jr., pp. 365–405. Ames, IA: Iowa State University Press. Noble, L. W. (1969). Fifty Years of Research on the Pink Bollworm in the United States, US Department of Agriculture Agricultural Handbook No. 357. Washington, DC: US Government Printing Office. Parencia, C. R. Jr. (1978). One Hundred Twenty Years of Research on Cotton Insects in the United States, US Department of Agriculture Agricultural Handbook No. 515. Washington, DC: US Government Printing Office. Perkins, J. H. (1980). Boll weevil eradication. Science, 207, 1044–1050. Pierce, D. L., Walters, M. L. Patel, A. J. & Swanson, S. P. (1995). Flight path analysis of sterile pink bollworm release using GPS and GIS. In Proceedings of the Beltwide Cotton Conference, 2, pp. 1059–1060. Memphis, TN: National Cotton Council. Pfadt, R. (1972). Fundamentals of Applied Entomology, 2nd edn. New York: Macmillan. Pfadt, R. E. (1978). Insect pests of cotton. In Fundamentals of Applied Entomology, 3rd edn, ed. R. E. Pfadt, pp. 369– 403. New York: Macmillan. Rabb, R. L. (1978). Eradication of plant pests: con. Bulletin of the Entomological Society of America, 24, 40–44. Staten, R. T., Walters, M., Roberson, R. & Birdsall, S. (1999). Area-wide management/maximum suppression of pink bollworm in Southern California. In Proceedings of the Beltwide Cotton Conference, pp. 985–988. Memphis, TN: National Cotton Council. Thomas, M. C., Heppner, J. B., Woodruff, R. E. et al. (2005). Mediterranean Fruit Fly. Originally published as DPI Entomology Circulars 4, 230 and 273, 2001; revised December 2005. Available at http://creatures. Walters, M. L., Staten, R.T., Roberson, R.C. & Tan, K. H. (2000). Pink bollworm integrated management using sterile insects under field trial conditions, Imperial Valley, California. In Area Wide Control of Fruit Flies and Other Insect Pests, Joint Proceedings of the International Conference on Area Wide Control of Insect Pests, May 28–June 2, 1998 and the Fifth International Symposium on Fruit Flies of Economic Importance, Penang, Malaysia, June 1–5, 1998., pp. 201– 206. Pulau Pinang, Malaysia: Penerbit Universiti Sains Malaysia. Weems, H. V. (1981). Mediterranean Fruit Fly, Ceratitis capitata (Wiedemann) (Diptera: Tephritidae), Entomology Circular No. 230. Tallahassee, FL: Florida Department of Agriculture and Consumer Services, Division of Plant Industry.

Chapter 24

Insect management with physical methods in pre- and post-harvest situations Charles Vincent, Phyllis G. Weintraub, Guy J. Hallman and Francis Fleurat-Lessard In theory, IPM programs should be an optimal blend of science (knowledge) and technologies – used concomitantly or sequentially – to manage pests below an economic injury level. There are five main approaches available to achieve that goal: chemical control (synthetic and naturally derived), biological control (predators, parasitoids and pathogens), cultural control (including cover crops and genetically resistant plants), physical control and human factors (legal restrictions on commodities, quarantines, etc.) (Vincent et al., 2003). In practice, few technologies are blended into most pest management programs. For both pre- and post-harvest pest control, the primary approach worldwide is chemical/fumigation (Fields & White, 2002). Like any technology, chemical control has its merits and limits; the development of resistance to pesticides by some arthropod populations, environmental contamination and tightening of regulations in registration and restrictions of use are among factors that limit the use of chemical control measures. However, human factors are playing an effective role in movement towards truly integrated control programs. For example, there have been new regulations enacted in North America (US Food and Drug Administration, 2004) and the European Economic Community (European Union, 2002) for hygienic food quality and safety (Table 24.1).

According to these new legislative measures, every food or feed product destined for trade must be free of arthropod pests. This requirement is the standard for food sanitary quality and hygiene of general application in international exchanges as established by the World Trade Organization (WTO). Physical control methods have been used for millennia. However, in the last two decades a revival of interest in alternative and sustainable pest control methods has prompted research projects in various agricultural contexts for both pre- and post-harvest pest control. Physical control encompasses: (1) an array of techniques (including applied engineering) and (2) knowledge from various scientific domains, notably physiology and population ecology.

24.1 Physical control techniques According to their mode of action, physical control methods can be active or passive (Table 24.2). The level of efficacy of active methods is proportional to both intensity of the energy and duration of its application to the target. Examples of active methods include thermal shock (heat, cold), electromagnetic radiations (microwaves, radio frequencies, infrared, ionizing radiations, UV and

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



Table 24.1

Recent European Union (EU) regulation dealing with the Food Quality and Safety Act

Scope of EU regulation Hygiene “package” Good processing and manufacturing hygiene practices Effective Hazard Analysis and Critical Control Point (HACCP) procedures application for good hygiene and wholesomeness of food Traceability of identity preserved products along the food chain

Official Regulations publication references Regulation 852/2004 on the hygiene of foodstuffs Regulation 178/2002: legislation about the respect of minimum hygiene requirements in food processing industries all along the food chain Regulation 178/2002: food business operators should establish and operate food safety programs and procedures based on HACCP and supported by the Guidelines for Good Hygiene Practices (GGHP) to enforce the regulation in all food chains Regulation 882/2004: implementation of official controls in every EU country to check food business operators compliance with respect to the minimum requirements to be provided for food safety

visible light), mechanical shock and pneumatic (blowing or vacuum). Most active methods have a short persistence as the stressor effect is limited to the period of application. This could be an advantage, as in post-harvest situations or situations where the environment should be minimally disturbed, or a disadvantage, in situations where the treatment must be repeated several times to achieve prolonged control. Passive methods do not require further energy to achieve desired effect. Examples are traps, airtight or hermetic storage, barriers of various kinds and trenches.

24.2 Pest physiology and population ecology Empirical use of physical control is acceptable but sustainable usage requires knowledge and this can only be accomplished through a thorough understanding of the biology and population dynamics of the pest. A case in point is the management of insects by using thermal sensitivity; one must determine the sensitivity of the pest and its substrate (either living or dead) to various temperatures to know whether thermal management is even feasible (Hallman & Denlinger, 1998). If an active physical control measure is to be used, knowledge of the population dynamics of the pest and the economic threshold

of the crop are required to know when to apply the measures.

24.3 Practical considerations In general, physical control methods are relatively more labor intensive and often time consuming when applied pre-harvest. The implementation of physical methods compares favorably with, and can augment, biological methods. However, most successes occur in post-harvest situations. Effective use of physical control measures, in both pre- and post-harvest control, relies on three main components: (1) good sanitation practices – which should be applied with all control methods, (2) monitoring to determine extent and stage of pest populations and (3) application of other direct or indirect physical control methods. An implementation scheme of these control measures for post-harvest protection is shown in Table 24.3. The effects of physical control methods (with the exception of electromagnetic radiation) are limited spatially. This attribute starkly contrasts with chemicals which may drift several hundred meters and may be bioaccumulated in food chains, or with biological control agents which may disperse actively or passively over long distances. Post-harvest IPM is greatly facilitated by the fact that it takes place inside a well-defined


Table 24.2

Classes, subclasses and key examples of physical control methods

Class (subclass) Passive Airtight or hermetic storage Aluminum foil Fences Flooding Inorganic (plastic) mulch Organic mulch

Packaging Screening Slippery surfaces

Key examples of target insectsa Stored product insects (A, L, P, E) Aphids (A) Anthomyiid flies (A), carrot rust fly (Psila rosae) (A) Cranberry insects, vegetable pests Tarnished plant bug (Lygus lineolaris) (L, A), anthomyiid flies Melonworm (Diaphania hyalinata), Colorado potato beetle (Leptinotarsa decemlineata) (E, L)

Tripple-bagging Windbreak

Insects of transformed products (A, L, P, E) Various insect species (A) Ants, cockroaches, Colorado potato beetle (A) Codling moth (Cydia pomonella) (L), tent caterpillars (L) Various dipteran pests – apple maggot fly (Rhagoletis pomonella) (A), Mediterranean fruit fly (Ceratitis capitata) (A), Colorado potato beetle (A) Chinch bug (Blissus leucopterus) (A), Colorado potato beetle (A) Cowpea insects (A, L, P, E) Aphid vectors (A)

Active Mechanical Dislodging Disturbing Forced air

Plum curculio (Conotrachelus nenuphar) (A) Stored product insects (A, L, P, E) (A, L)

Inert dusts

Stored product insects (L, A)

Infrasound Mechanical impacts Mechanical polishing Mineral and vegetable oils Overhead irrigation Particle films (kaolin) Physical removal of hosts Plant residue removal Pneumatic

Various insect species Stored product insects (A, L, P, E) Rice weevil (Sitophilus oryzae) Phytophagous mites, soft bodied insects Diamondback moth (Plutella xylostella) Codling moth, leafrollers, mites (L) Apple insects (A) Millet stem borer (Coniesta ignefusalis) (L, P) Colorado potato beetle (A), Lygus spp. (L, A), Obliquebanded leafroller (Choristoneura rosaceana) (L)

Sticky barriers Trapping



Context and commentsb PH Pr; repels aphids Pr Pr; cranberry In strawberry Pr; indirect effect on natural enemies, positive interaction with Bt PH Greenhouse, orchards Pr; made of fluon, teflon or dust Pr Pr; fruit orchards

Pr PH Pr

Pr; in apple orchards Pr; fluidized bed PH; increase insecticide effect PH; including diatomaceous earth, silica dust Pr PH, Entoleter In rice Pr, tree fruits Pr Pr Removal of host plants Pr; vacuuming, blowing Pr (cont.)




Table 24.2


Class (subclass) Pulsed ultrasound Sieving Surface tension agents and sufactrants Ultrasound Vacuumized packaging Thermal Burning Flaming High temperatures Hot water–steam Infrared heating Low temperatures Post-harvest chilling Rapid freezing Solar heating Steam Electromagnetic Ionizing radiation Microwave Radio frequencies UV and visible light Modified atmospheres CO2 Inert gases N Plastic sheeting

Key examples of target insectsa

Context and commentsb

Various insect species Insects of peanuts Phytophagous mites (A, L)

Pr PH Pr; soaps

Various insect species Weevils

Pr Polyester film bags

Maize stalk borer (Busseola fusca) Colorado potato beetle (A) Stored product insects Fruit (E, L)

PH; partial stalk burning Pr PH PH PH PH PH PH PH; stored grains PH; potato

Stored product insects Stored product insects Stored product insects Cowpea weevil (Callosobruchus maculatus) Colorado potato beetle (A) Stored product insects Stored product insects Stored product insects Stored product insects

Stored product insects Stored product insects




A, adult; E, eggs; L, larvae; P, pupae. PH, post-harvest; Pr, pre-harvest. Source: Adapted from Supplemental material of Vincent et al. (2003). Reprinted, with permission, from c 2003 by Annual Reviews. the Annual Review of Entomology, 48  b

area, delimited by the structures of industrial buildings (food processing facility, grain storage elevator, food factory, feed mill, etc.). As a consequence of all of these aspects, physical control methods occupy a smaller market share than pesticides worldwide. For further information on physical control methods in agricultural plant protection, we refer the reader to the reviews of Banks (1976), Oseto (2000), Vincent et al. (2001, 2003) and Weintraub & Berlinger (2004).

24.4 Pre-harvest control measures 24.4.1 Exclusion barriers Insect exclusion screening is probably the single most important physical control method developed in the last century. Following the invasion of virus-bearing whiteflies, tomato crops in the entire Mediterranean region could not be


Table 24.3 Development of an IPM system as a part of the HACCP procedures preferably using physical preventative and control means: effective IPM based on biophysical methods applicable in post-harvest situations Inspection and identification of infestation risks on plant structure and material

Detection, surveillance and monitoring systems


Prevision of infestation risk by pest population dynamics models

Implementation of preventive measures and control vs pre-established limits


Species identification

UV light traps

Simulation of population kinetics Specific monitoring measures

Cleaning and sanitation of surroundings and inside plant Manipulation of physical conditions Mass-trapping

Pheromone traps

Application of corrective treatments for the limitation of damage Heat disinfestation

Use of a decision support system based on expertise

Temporary freezing

Replanning of optimized strategies Training the personnel in charge of IPM strategy application Information of employees about their implication into the IPM system

Microwave or radiofrequency heating

Complementary Modified measures for atmosphere pest exclusion packaging from food plant by physical barriers Controlled atmosphere storage Cold storage Combination of physical treatments

grown in open fields from late spring through fall. The development of screens allowed tomatoes to be produced year-round, and the use of screens has become a standard pest management practice worldwide. This form of physical control has proven cost-effective both for consumers and growers (Weintraub & Berlinger, 2004). Usage of exclusion screening is not limited to the traditional greenhouse crops; orchards are increasingly being covered to prevent pests from gaining access to trees. Netting which excludes

Replanning the prevention plan from new data or experience

fruit flies has proven to be an economically effective method of protecting peaches and nectarines, increasing yield quantity and quality (Nissen et al., 2005a, b). Certain diseases, such as papaya dieback, can only be controlled by covering plantations with coarse white nets to limit vector movement (Franck & Bar-Joseph, 1992). Bananas are grown under screening in many parts of the world. Another example of an exclusion barrier in northeastern North America involves the simultaneous management of weeds in apple orchards




and two insect pests, plum curculio (Conotrachelus nenuphar) and apple sawfly (Hoplocampa testudinea) (Benoit et al., 2006). Both insects lay their eggs in fruitlets where the larvae develop. In late June infested fruitlets fall to the ground and mature larvae enter the soil to pupate. In field experiments over a four-year period, cellulose sheeting prevented most weeds from emerging (Benoit et al., 2006). Total weed density was significantly lower in plots covered with cellulose sheeting as compared to control (no sheeting) plots. However accumulation of soil on the sheeting allowed a few weeds to grow. Likewise, emergence of adults from fallen apples covered with a cage was significantly reduced for both plum curculio and apple sawfly. However, some individuals completed their development and successfully overwintered on the sheeting. For example, in spite of cellulose sheeting, the percentage of plum curculio emergence varied from 0.7% to 18.9% in one orchard compared to 9.3% to 63.5% in the control. Over the years, the integrity of cellulose sheeting has been challenged by accumulation of plant debris and water, causing biodegradation of the sheeting, allowing weeds and insects to penetrate through damaged surfaces. The potential effect of exclusion barriers on weeds depends on the size of the seed bank at the beginning and the viability of the seeds. If one uses a resistant material, these results imply that over medium to long term (five to ten years), populations of both insects would decline in absence of nearby (1.5 million bales of cotton valued at $US 1250 million in yield reduction and control costs (Williams, 2006). As will be discussed below, many factors have contributed to reductions in pest losses over the past 20 years including boll weevil eradication, transgenic cottons for control of caterpillar pests and improved overall IPM programs for various pests.

25.1.2 Beneficial arthropods Many beneficial arthropod species are associated with cotton. In a seminal study, Whitcomb & Bell (1964) cataloged about 600 species of arthropod predators including ∼160 species of spiders in Arkansas cotton fields. Van den Bosch & Hagan (1966) suggested that there may be nearly 300 species of parasitoids and arthropod predators in western cotton systems. Like cotton herbivores, only a fraction of these are common, with some of the more abundant groups including big-eyed bugs (Geocoris spp.), anthocorid bugs (Orius spp.), damsel bugs (Nabis spp.), assassin bugs (Zelus and Sinea spp.), green lacewings (Chrysopa and Chrysoperla spp.), lady beetles (e.g. Hippodamia and Scymnus spp.), ants (especially Solenopsis spp.), a wide variety of web-building and wandering spiders and both parasitic wasps (e.g. Bracon spp., Cotesia spp., Microplitis spp., Hyposoter spp., Trichogramma spp.) and flies (Archytas spp., Eucelatoria spp.). We continue to learn about the important role that


these natural enemies play in cotton pest control but the most dramatic evidence of their impact comes from studies in which the destruction or disturbance of natural enemy communities by indiscriminant insecticide use is associated with pest outbreaks (e.g. Leigh et al., 1966; Eveleens et al., 1973; Stoltz & Stern, 1978; Trichilo & Wilson, 1993). Overall, arthropod communities in cotton are dynamic and largely driven by the wide diversity of management options discussed below.

25.2 IPM tactics 25.2.1 Chemical control Insecticide use has a long history in USA cotton pest control beginning with arsenicals for control of boll weevil in the early twentieth century and followed by a progression of synthetic insecticides (e.g. organochlorines, organophosphates, carbamates and pyrethroids) in the subsequent decades following World War II (Herzog et al., 1996; Sparks, 1996). With each new introduction, periods of excellent pest control were generally followed by control failures due to the evolution of insecticide resistance. Many past and current insecticides have broad activity against pests and their associated natural enemies and pose hazards to the environment and human health. On a worldwide basis cotton accounts for about 22.5% of all insecticide use (Anonymous, 1995) and historically USA cotton has been the heaviest user of insecticides. That pattern has begun to shift (Fig. 25.3) with the introduction of transgenic Bt cotton in 1996 and the availability of a wide range of effective and safer insecticides registered in part through the US Environmental Protection Agency (EPA)’s Reduced-Risk Initiative over the past decade (Environmental Protection Agency, 2006) (Table 25.2). While a variety of many classes of insecticides continue to be used for cotton pest control throughout the USA (National Agricultural Statistics Service, 2006), adoption and use of reduced-risk insecticides has grown in recent years (e.g. Goodell et al., 2006).

Fig. 25.3 Overall insecticide use patterns in USA cotton 1986–2005 summarized from Cotton Insect Losses, a database compiled by the National Cotton Council (2007b).

implementation of chemical control and careful vigilance and proactive strategies are needed to preserve this important tactic (Castle et al., 1999) (see Chapter 15). Around 550 arthropod pests have developed resistance to one or more insecticides, and currently a total of 34 cotton pests (19 in the USA) have developed resistance to as many as three insecticide classes (Whalon et al., 2007). The mitigation of resistance is based on management of insecticide type and use that either attempts to reduce the fitness of resistant individuals or minimizes selection pressure on a pest population (Roush & Daly, 1990). Simply put, this means limiting insecticide use through adherence to economic thresholds, diversifying modes of action through rotations, mixtures and use of synergist, and partitioning of insecticide use in space and time by adoption of seasonal stages or cropspecific usage. Examples include the Australian IRM strategy for managing resistance in the Old World bollworm (Heliocoverpa armigera) through a three-stage plan which rotates pyrethroids with non-pyrethroids over the season (see Castle et al., 1999), the Texas and Midsouth pyrethroid use window strategy for resistance in tobacco budworm (Plapp et al., 1990) and the multi-crop resistance management plan for whitefly in the western USA in which various classes of insecticides are rotated depending on predominant crop mixtures within a region (Palumbo et al., 2001, 2003).

25.2.3 Cultural control 25.2.2 Resistance management The development of resistance in pest populations to insecticides is a continual threat to successful

The indeterminate nature of cotton plant growth and the influence of production practices such as cultivation, irrigation, fertilization, cultivar




Table 25.2

Insecticides registered for use on cotton through the EPA Reduced-Risk/Organophosphate Alterna-

tives Program


Mode of action (IRAC MoA)a

Spinosad Pyriproxyfen Tebufenozide Methoxyfenozide Indoxacarb Thiamethoxam Buprofezin Pymetrozine Bifenazate Acetamiprid Etoxazole Novaluron

Acetylcholine receptor modulator (5) Juvenile hormone mimic (7) Molting hormone agonist (18) Molting hormone agonist (18) Sodium channel agonist (22) Acetylcholine receptor agonist (4A) Chitin synthesis inhibitor (16) Feeding blocker (9) Neural inhibitor (25) Acetylcholine receptor agonist (4A) Growth inhibitor (10B) Chitin synthesis inhibitor (15)

Fenpyroximate Dinotefuran Flonicamid Spiromesifen

Electron transport inhibitors (21) Acetylcholine receptor agonist (4A) Feeding blocker (9) Lipid synthesis inhibitor (23)

Cotton target

Year registered

Caterpillars Whiteflies, aphids Caterpillars Caterpillars Caterpillars Whiteflies, aphids, thrips Whiteflies, aphids Whiteflies, aphids Mites Whiteflies, aphids, plant bugs Mites Whiteflies, thrips, caterpillars, plant bugs Mites Whiteflies, thrips, plant bugs Aphids, plant bugs Whiteflies, mites

1997 1998 1999 2000 2000 2000/2001 2001 2001 2002 2002 2003 2004 2004 2005 2005 2006

IRAC MoA = Insecticide Resistance Action Committee Insecticide Mode of Action Classification (IRAC, 2007).


selection, weed control and planting date on the crop’s susceptibility to damage and suitability for insect infestations remain an extremely important aspect of effective pest management (Ridgway et al., 1984; El-Zik et al., 1989; Matthews, 1994; Walker & Smith, 1996). For example, stalk destruction, field sanitation, efficient harvest, tillage and winter irrigation can effectively control or reduce populations of boll weevil and pink bollworm (Walker & Smith, 1996). Early planting and early crop termination are long-standing principles of cultural control and pest avoidance that are still relevant for many pest species. From eastern Texas to the Atlantic coast, timely planting and early harvest helps to avoid fall and winter rains and resulting in important economic advantages (Parvin & Smith, 1996). Delayed planting also may have benefits. For example, planting later so that no fruiting forms are present when pink bollworm adults emerge from the soil maximizes “suicidal emergence” and reduces pest populations throughout the season (Brown et al., 1992). Cultivation can effectively reduce overwintering

survival of bollworm and tobacco budworm, and Schneider (2003) suggested that recent trends for reduced tillage could accelerate resistance of tobacco budworm to insecticides and Bt toxins in transgenic cotton (see below). Deep tillage induces high overwintering mortality in pink bollworm (Watson, 1980). There has been renewed interest in manipulating dispersal and crop colonization though trap crops, especially in the management of polyphagous plant bugs. The use of strip crops of alfalfa within cotton is the classic example of cultural control through the practical deployment of trap crops (Stern et al., 1964). Most examples of trap cropping, including this classic example of strip harvesting alfalfa, have been only sporadically accepted and utilized in production agriculture because of the logistic impact on farming operations and the wide availability of effective chemical controls options (Shelton & BadenesPerez, 2006). For example, only 4% of California growers reported using manipulation of alfalfa to control cotton pest insects (Brodt et al., 2007).


Future management systems may need to examine incentives for grower adoption and expansion of cultural management tactics that may reduce pest populations across broad geographic regions.

25.2.4 Behavioral control A suite of tactics are available that alter or manipulate the behavior of pest arthropods leading to population suppression or even elimination. Two examples, (1) pheromones and (2) sterile insect release, are discussed here. Pheromones A pheromone is a chemical that mediates behavioral interactions between members of the same species. The most common examples are sex pheromones which are involved in mating, but aggregation and alarm pheromones are also known from cotton pest species. As of 1994 sex pheromones have been identified in 15 major cotton pest species (seven species in the USA) including 14 moths and one beetle (Campion, 1994). The three main applications of pheromones are monitoring, mating disruption and mass trapping (see Chapter 21). Traps baited with sex pheromones are routinely used for selective monitoring of pink bollworm, boll weevils, bollworm and tobacco budworm. Trapping information is useful in pest detection at low densities and tracking seasonal events such as adult emergence and the number and timing of generations. For pink bollworm, pheromone traps have even been used to monitor density for pest control decision making (Toscano et al., 1974) and traps are a major component of the ongoing pink bollworm eradication program (see below). Pheromone traps continue to play an important role in ongoing boll weevil eradication efforts and as long-term monitoring tools in posteradication areas of the USA. Mating disruption is achieved by applying pheromone to a field, thereby making it difficult for potential mates to find one another and resulting in reduced mating and subsequent reproduction. This technology was first applied in 1978 for the pink bollworm and the method is still used today. Mating disruption is a major component of the ongoing eradication and exclusion

programs for this pest and has been used in several past areawide programs in California and Arizona and in other countries (Campion, 1994). Mating disruption has been evaluated for other pests such as bollworm and tobacco budworm, but their polyphagous nature is problematic and results have been unsuccessful or ambiguous (Campion, 1994). Mass trapping showed some promise for pink bollworm control in a three-year grower funded trial in Arizona (Huber et al., 1979) and the method was used in some Arizona production areas into the mid-1990s. The approach is not currently in use for pest control in cotton. Sterile insect release The notion that mass release of sterile insects could be used to manage or eradicate a pest was first proposed by E. F. Knipling during the 1930s (Knipling, 1955) and was first successfully used to eradicate screwworm (Cochliomyia hominivorax) from the island of Curac¸ao during the 1950s (Baumhover et al., 1955). The concept, known as the sterile insect release method (SIRM) or the sterile insect technique (SIT) has been attempted with several major cotton pests in the USA including the bollworm and tobacco budworm, the boll weevil, and most successfully with the pink bollworm. Various biological and operational factors precluded the successful application of SIRM to the two former species/groups (see Villavaso et al., 1996) but the method has been used annually since 1968 to help mitigate the establishment of pink bollworm on cotton in the Central Valley of California (Miller et al., 2000), and is a component of the current pink bollworm eradication program.

25.2.5 Host plant resistance Host plant resistance is a fundamental management tactic (El-Zik & Thaxton, 1989; Gannaway, 1994; Jenkins & Wilson, 1996). Host plant resistance can be broadly categorized as antibiosis (reduced fitness or pest status), antixenosis (avoidance or behavioral factors) and tolerance (ability of plant to compensate for damage) (see Chapter 18). Plant resistance traits may include manipulation of the plant’s genome or the resistance may be associated with indirect selections for




traits like yield and fiber quality. Genetically controlled traits useful in cotton resistance to insects include: crop earliness, a range of plant morphological traits (nectariless, glaborous or pilose leaf surface, okra-shaped leaf, frego bract, red plant color, yellow or orange pollen) and varying concentrations of plant secondary compounds (high gossypol and tannin content). Relatively few of these traits have been incorporated into commercial cultivars. The tools of biotechnology have provided new opportunities to enhance traditional approaches to host plant resistance (see Chapters 18 and 21). The impact of transgenic cottons producing insecticidal toxins from Bacillus thuringiensis (Bt), primarily for control of pink bollworm, tobacco budworm and bollworm has been enormous. The reduction in insecticide use in cotton over the past decade (Fig. 25.3) can be partly attributed to increased adoption of Bt cottons. Use of Bt varieties has expanded each year with 57% of all USA upland cotton hectares planted to Bt cotton in 2006 (Williams, 2007). In most areas of the Midsouth and Southeast where bollworm and tobacco budworm are historically important pests, adoption of Bt cotton approaches 80–90%. In 2006, Texas planted about 35% of its total cotton hectares to Bt varieties while California planted less than 20% (a large portion of California’s hectares are planted to long staple Pima varieties that have not been transformed). In Arizona, Bt cottons are widely adopted because of their dramatic impact on pink bollworm. Commercial transgenic cottons with insecticidal activity are currently limited to the transgenes from B. thuringiensis. Bollgard cotton (Monsanto Company, St. Louis, MO) was the first commercially available cotton in 1996. It expresses the Cry1Ac insecticidal protein. Bollgard II cotton (Monsanto) expressing Cry1Ac and Cry2Ab2 protein was commercially introduced in 2003, and Widestrike cotton (Dow AgroSciences, Indianapolis, IN) expressing Cry1Ac and Cry1F protein was launched in 2005. VipCot cotton (Syngenta Biotech, Jealott Hill, Berkshire, UK) will express the Vip3A vegetative protein from B. thuringiensis, probably along with a Cry protein, and is expected to be commercialized in the USA shortly.

One of the most hotly debated issues facing cotton insect management is the sustainability of transgenic Bt cottons. However, despite the high use of Bt crops there has been little or no increase in insect resistance over the ten years of commercial deployment (Tabashnik et al., 2003). This success can be partly attributed to a EPAmandated resistance management program that requires growers using Bt cotton to also plant non-Bt cotton refuges. The principle behind this mandated strategy is that non-Bt cottons produce susceptible target pests that can readily interbreed with any resistant pests that may arise from Bt fields, thereby diluting incipient resistant populations. Bt cottons offer real environmental and economic advantages to conventional cottons sprayed more with insecticides. Frisvold et al. (2006) estimated global economic benefits of $US 836 million for Bt cotton in USA. The potential role of Bt cotton in reducing human exposure to toxic chemicals, especially in developing countries where insecticides are often applied manually, is large. Still, environmental risk issues such as effects on non-target organisms and ecosystem function and gene flow associated with transgenic crops in general continue to be debated and researched in the scientific community (e.g. Andow et al., 2006; Romeis et al., 2006). There has been limited progress in the development of transgenic cottons for pests other than caterpillars. For example, Monsanto is in the very early stages of development of transgenic cottons targeting Lygus spp. based on Bt and non-Bt approaches. Focus on non-caterpillar pests remains a major goal of various biotech firms and basic researchers around the world, and it highlights the importance of a continued investment in traditional host plant resistance.

25.2.6 Biological control The three major approaches to biological control include classical biological control, where exotic agents are introduced for permanent establishment against exotic and native pests, augmentation biological control, which involves the rearing and periodic release of natural enemies, and conservation biological control, which attempts to protect, manipulate and enhance


existing natural enemies for improved control (see Chapters 9 and 12). Classical biological control programs have been carried out in the past for several pest groups including bollworm, tobacco budworm, boll weevil, pink bollworm and to a lesser extent for lygus bugs. These efforts have been largely unsuccessful in the cotton system as natural enemies have either failed to become established and/or their impacts have been minimal (King et al., 1996a). The whitefly B. tabaci has been the most recent target of classical biological control, with numerous species of parasitoids released for establishment (Gould et al., 2008); however, as with other classical efforts, the impact of these established agents have so far been minimal in the cotton system (Naranjo, 2007). Likewise, augmentative biological control with predators and parasitoids has been researched and evaluated for several major cotton insect pests, but factors such as lack of efficacy, technical difficulties with natural enemy massproduction and cost relative to insecticides have combined to limit this approach from becoming a viable option in cotton pest control in the USA (King et al., 1996a). Augmentation with microbial agents (viruses, fungi, bacteria) for control of bollworm, tobacco budworm, boll weevil, pink bollworm, whitefly and plant bugs (King et al., 1996a; Faria & Wraight, 2001; McGuire et al., 2006) has been examined; however, commercialized microbial products continue to have very small shares of the cotton pest control market. In contrast to classical and augmentative biological control, conservation biological control continues to be a major focal area of cotton IPM that has been further stimulated by the many recent changes to cotton pest management systems. As noted, the cotton system in the USA harbors a diverse complex of native natural enemies, many of which are generalist feeders that opportunistically attack many insect and mite pests. Naturally occurring epizootics of some microbes also may significantly suppress pest species (Steinkraus et al., 1995). The potential value


of these natural enemies in pest suppression has been repeatedly demonstrated over many decades in the cotton system when broad-spectrum insecticides applied for one pest lead to resurgence of the target pest and/or outbreaks of secondary pests through the destruction of natural enemies (see Bottrell & Adkisson, 1977). This potential is also widely recognized in state recommendations for cotton IPM. Most guidelines call for sampling of natural enemies and emphasize their preservation through inaction or judicious use of insecticides, particularly those with selective action. Our understanding of the role and interaction of natural enemy species and complexes and how to manipulate them for improved pest control in cotton has a rich history that continues to grow (e.g. Sterling et al., 1989; Naranjo & Hagler, 1998; Prasifka et al., 2004).

25.2.7 Sampling and economic thresholds A hallmark of all cotton IPM programs in the USA is monitoring of pest density or incidence combined with action or economic threshold to determine the need for control measures. In 2006, about 50% of USA cotton hectarage was scouted an average of 1.3 times per week at an average cost of $US 18.97/ha across the cotton belt (Williams, 2007). The intensity of scouting varies greatly by state. In Virginia and Kansas less than 5% of cotton hectarage was scouted but more than 90% was scouted in Arizona, Louisiana and South Carolina. In Texas, the largest cotton producing state, only 19% of the hectarage was reported as scouted. The cooperative extension programs of each of the 17 cotton growing states produce recommendations for scouting, treatment thresholds and insecticides to help growers and consultants implement IPM.1 The basic tools of sampling include sweep nets, beat cloths and beat boxes, traps and visual inspection of various plant parts, all of which require human labor and the associated cost. Sampling plans, which specify the general protocols for how samples should be collected and how many sample units should be taken, are

Web links to sampling and threshold recommendations and individual state recommendations are available at http://ipmworld.




typically based on research to understand the distribution and variability of pest populations (see Chapter 7). Sequential sampling plans, which minimize the number of sample units that need to be taken, are often developed for cotton pest management application, but in practice it is more typical for a set sample size to be recommended and implemented. Many cotton pest sampling plans also use a presence/absence approach (e.g. percent infested) for monitoring rather than a complete count which allows for quicker sampling and decision making. Thresholds tend to vary depending on production region and also may be dynamic, with critical pest densities being a function of plant development and prior management activity. For example, thresholds for bollworm and tobacco budworm are lower following an initial insecticide application during post flower bloom, while thresholds for plant bugs generally increase as the cycle of flower bud (square) production progresses. Thresholds in Bt cotton also may differ from non-Bt cotton for the bollworm which is not controlled completely in some transgenic cottons. Some of these thresholds are based on experimental study (see Chapter 3), but many are nominal thresholds developed on the basis of trial-and-error experience by researchers, extension agents, consultants and growers. Many state guidelines encourage scouting of natural enemy populations, but only a few have provided explicit information on how to use these counts to modify treatment decisions (e.g. Fillman & Sterling, 1985; Wilson et al., 1985; Conway et al., 2006). Nonetheless, the preservation of natural enemies through judicious use of insecticides is implicitly recognized as a key component of most management systems.

logical and economic factors (Mumford & Norton, 1994; Kogan, 1998) (see Chapter 38). Education is a key element in IPM implementation regardless of the crop and Cooperative Extension Services associated with land-grant universities generally take the lead in developing educational materials (circulars, bulletins, websites) as well as organizing training and even on-farm demonstrations and adaptive research. Private industry may also contribute educational and consulting activities, and depending on the scope of implementation, grower organizations and/or federal agencies such as USDA may be involved (Harris et al., 1996).

25.3 IPM programs and implementation

25.3.2 Models and decision aids

Despite the challenge of researching and compiling the necessary component tactics into a workable IPM strategy, the greatest difficulty may be in implementing and evaluating IPM programs because such tasks depend on logistical, socio-

25.3.1 Areawide programs Cotton entomologists have long recognized the importance of spatial scale of management activities for mobile pest species. Ewing & Parencia (1950) demonstrated effective control of boll weevil when coordinated early-season treatments were applied on a community basis. With the dramatic success of screwworm eradication, Knipling (1979) extended the areawide concepts of pest population suppression to cotton insects, especially the boll weevil and the tobacco budworm. Henneberry & Phillips (1996) provide an overview of the elaborate experiments and theoretical debates that followed. Boll weevil eradication (see below) is perhaps the largest-scale example of areawide programs. Other examples include numerous attempts to manage bollworm and tobacco budworm through biological control and community management systems, a 40-year program to exclude pink bollworm from central California, successful control of whitefly in the desert valleys of the West, and emerging management systems for plant bugs and stink bugs in the Midsouth and Southeast.

A wide diversity of simulation models and decision support systems have been developed for USA cotton beginning in the 1970s with the NSF/EPA Integrated Pest Management Project (commonly known as the “Huffaker project”), continuing with the USDA/EPA Consortium for IPM project during the 1980s and many other cooperative state and federal projects thereafter (Mumford & Norton,


1994; Wagner et al., 1996). Modeling efforts have focused on individual pests and groups of pests and some have included simple or detailed cotton plant models. In general, these models have been useful in structuring knowledge of the plant–pest system, studying and predicting population dynamics, examining alternative management scenarios and identifying areas needing further research. However, with few exceptions such models have found very limited application in guiding day-to-day pest and crop management activities. In the 1990s, management tools based on expert systems and information management began to be developed for cotton insect management (Wagner et al., 1996). Some of these decision aids were coupled with the more complex simulation models and others were based on broad generalizations of expert opinion and historical data. These systems included the expert systems rbWHIMS, COTFLEX, CALEX and CIC-EM for a range of cotton production systems. In general, none of these systems has seen wide-scale adoption. COTMAN, a more recent decision aid, emphasizes the synthesis of field samples rather than projection of information and has remained a component of practical cotton production in limited areas of the Midsouth, Southeast and Texas. The strength of COTMAN and its continued use may be due to the simplicity of the system and its close conceptual linkage to crop development.

25.3.3 Case studies There are many examples of operational IPM programs for cotton pests throughout the USA that involve many of the component tactics discussed. Here we highlight two representative programs, (1) whiteflies in Arizona and (2) plant bug and stink bug management in the Midsouth and Southeast. Whitefly IPM strategy in Arizona Since the early 1990s the polyphagous sweetpotato whitefly (B. tabaci), Biotype B, has had major impacts on most agricultural production in the West (Oliveira et al., 2001). In response, a multi-component research and educational plan was launched that resulted in a successful IPM

program which continues to be expanded and refined today (Ellsworth & Martinez-Carrillo, 2001; Naranjo, 2001; Ellsworth et al., 2006). The overall program can be envisioned, and is taught to growers and consultants, as a pyramid with multiple, overlapping layers and components (Fig. 25.4). The broad base of the pyramid, founded on research, emphasizes tactics and strategies that can be implemented to reduce overall pest populations including various crop management practices and selection of well-adapted, smooth-leaf varieties which are generally less attractive to whiteflies. The foundation also emphasizes natural enemy conservation through the use of selective control methods for whitefly and other pests and an array of areawide tactics like crop placement and arrangement to reduce pest movement, destruction of crop residue and weeds and coordinated use of insecticides among all affected crops to manage resistance. The two upper layers of the pyramid outline pest monitoring through an efficient binomial sampling scheme (Ellsworth et al., 1996; Naranjo et al., 1996), and the timing of effective control methods based on economic threshold and a three-stage insecticide use system which emphasizes selectivity (i.e. safety to beneficial arthropods) in the initial stages (Naranjo et al., 2004; Ellsworth et al., 2006). Follow-up treatments are rarely needed if these selective options are used first because the conserved natural enemies and other natural forces are then able to suppress whitefly populations long term (Naranjo, 2001). The three-stage system also implicitly encourages the rotation of insecticides with differing modes of action in order to mitigate resistance. Operationally, the IPM plan has significantly reduced insecticide use for all cotton pests in Arizona from a decades-long high of over 12 applications in 1995 at a cost of $US 536/ha to a decades low application rate of 1.4 at a cost of $US 77/ha in 2006.

Plant bug and stink bug management in the Midsouth and Southeast A plant bug complex, including tarnished plant bugs (Lygus lineolaris), cotton fleahoppers (Pseudatomoscelis seriatus) and clouded plant bugs




Fig. 25.4 Conceptual diagram of the Arizona Whitefly (WF) IPM program showing the three main components of effective IPM: avoidance, effective use of insecticides and sampling. (Reprinted from Ellsworth & Martinez-Carrillo [2001] with permission from Elsevier.)

(Neurocolpus nubilis), has been a long-standing pest problem in Midsouth and Southeast cotton. Plant bugs typically attack cotton at early squaring but can persist as a pest problem through boll development. A complex of seed-feeding stink bugs (brown stink bug [Euschistus servus], green stink bug [Acrosternum hilare] and southern green stink bug [Nezara viridula]) also attack maturing bolls later in the growing season. Both plant bugs and stink bugs are increasing in status but these tend to be more important in the Midsouth and Southeast, respectively. The elevated importance of these polyphagous and mobile insects reflects success in eliminating boll weevil (through eradication) and tobacco budworm (by Bt cotton) as major pests. In 2006, crop loss and insecticide use for these bug pests were twice to three-fold those of other pests across the Midsouth and Southeast (Williams, 2007). Designing effective control measures for the bug complex has been a challenging and difficult task. Plant bugs are now resistant to several insecticides (Snodgrass, 1996). The USDA in Stoneville,

Mississippi has developed an areawide approach to the removal of early-season broadleaf hosts of tarnished plant bug (Snodgrass et al., 2005). This research approach has been evaluated by extension entomologists across the Midsouth and is being adopted on limited hectares by growers in some regions. Additional testing is needed to confirm the broader impacts on other pest and beneficial species in the system. Extension entomologists in the Southeast are developing treatment thresholds and monitoring procedures for the stink bugs (Greene et al., 2001), and those in the Midsouth are studying sampling and management options for plant bugs.

25.4 Pest eradication Pest eradication involves the complete elimination of a pest from its current range and generally focuses on invasive pest species (Chapter 10). Significant debate continues over the value of pest eradication as a substitute for or as a complement to pest management (Myers et al., 1998). Nonetheless, eradication programs are currently under way for two key cotton pests in the USA and they will be briefly discussed here.


25.4.1 Boll weevil Effective removal of boll weevil as a key pest of USA cotton is an important biological and social achievement covering a half-century of scientific and strategic effort (Cross, 1973; Smith & Harris, 1994; Dickerson et al., 2001; Hardee & Harris, 2003). Interest in a coordinated effort to eliminate this invasive cotton specialist began in the late 1950s and the eradication program was initiated in earnest by the late 1970s. Incipient populations were eradicated in Arizona, California and northwest Mexico in the 1980s, and the nationwide effort began in North Carolina and successfully expanded through South Carolina, Georgia, Florida and south Alabama. In the early to mid1990s boll weevil was eradicated from central and north Alabama, middle Tennessee and the Texas Southern Rolling Plains. During the late 1990s, boll weevil eradication expanded to the Midsouth and Texas, with only a few isolated regions still infested. It is anticipated that the USA will be weevil free by 2009. Reductions in cotton insect losses (Table 25.1) can be directly and indirectly related to removal of this key USA cotton pest. In the past, early-season treatments were necessary to keep weevil populations from expanding, but they also triggered many additional pest outbreaks. The basic components of the program include monitoring with pheromone (grandlure)-baited traps to time early-season applications of insecticides that reduce establishment in cotton, late-season “diapause” treatments to reduce overwintering weevils and early crop maturity and crop destruction to enable a host-free period coordinated across large geographic zones.

25.4.2 Pink bollworm Like the boll weevil, the pink bollworm is an exotic cotton specialist that successfully invaded in USA in the early 1900s and became firmly established in the West by the mid 1960s following various attempts to contain and suppress populations throughout the first half of the twentieth century (Henneberry & Naranjo, 1998). The cooperative eradication program involves growers, and state and federal agencies. The program is being implemented in phases beginning with west Texas, New Mexico and northern Chihuahua, Mexico in 2001 and continuing through Arizona to southern

California and northern Sonora, Mexico in 2007 with the total program completed by 2010. The basic elements of the program include mapping and monitoring of all cotton fields within each region and the use of a combination of Bt transgenic cotton, mating disruption with pheromones, sterile insect release, and follow-up insecticides as needed. The sterile insect release in this case serves both to augment population control and as a substitute for the required nonBt refuge for resistance management which was relaxed in Arizona and southern California to allow for 100% production of Bt cotton. Pink bollworm populations in the Phase I regions have been reduced by >99% from 2001 to 2005 (El-Lissy & Grefenstette, 2006), but it is too early to gauge the overall success of the eradication effort.

25.5 Conclusions IPM is based on an ever-changing foundation of improved scientific knowledge, economic circumstances, and societal issues and demands. Several significant technological advances (e.g. transgenic crops) have occurred in the past decade that have dramatically lowered pest losses and significantly lowered insecticide use in a system that has historically been associated with insecticide over-reliance and misuse. Undoubtedly, future advancements will continue to improve the sustainability and environmental quality of cotton production in the USA and worldwide. Environmental issues will continue to grow, especially as urban areas expand and become more closely integrated with crop production areas. Current IPM programs in cotton like many other crop systems are largely focused on what Kogan (1998) characterizes as “Level 1 IPM,” or IPM of single pest species in individual fields. This is contrasted to “Level 2 IPM” which focuses on interactive effects of multiple pest species within whole farms or “Level 3 IPM” which involves management of multiple pests on perhaps multiple crops within entire agroecosystems. Some of the areawide programs summarized above have begun to view and manage cotton pests within a broader landscape perspective, but much additional research will be needed to understand the simultaneous and multiple impact of all pests




(insects, weeds, pathogens) on plant health and to develop efficient decision aids and control methods for managing multiple stressors at multiple spatial scales. This task will be an even greater challenge for polyphagous and mobile pests such as aphids, mites, whiteflies, plant bugs, bollworm and tobacco budworms. Meeting these challenges will likely call upon the increased use of models, risk assessment tools and information technology at both the grower and regulatory level to better understand, predict and manage systems behavior. This is going to require information managers and more userfriendly systems for storing, mining, analyzing and applying this information to farm-level decisions. Transgenic cotton conferring either insect or herbicide resistance or both has been widely adopted by growers to manage risk from caterpillar and weed pests and that trend is likely to continue. Current commercial cultivars of insecticidal transgenic cotton are based on one or more of three Cry toxins and one vegetative insecticidal protein but other proteins like snowdrop lectin (GNA) and protease inhibitors are being examined in other crop species (Christou et al., 2006), and over 170 distinct δ-endotoxins as well as many other toxins are known from B. thuringiensis (Glare & O’Callaghan, 2000) providing much to be mined for future transgenic plant development targeting multiple pests. The technologies associated with precision agriculture (GIS, remote sensing and GPS) are likely to expand (Shaw & Willers, 2006). Such technologies may reduce overall inputs like fertilizers, herbicides and insecticides by selectively allowing growers to apply these only as needed within specific areas of a single field. Overall, the long tradition of pest management research and practice in the cotton system will continue to expand, leading to reduced risk and greater predictability for producers, and greater sustainability and environmental stewardship benefiting society as a whole.

Acknowledgements We thank Drs. Pete Goodell (University of California), Peter Ellsworth (University of Arizona) and John Ruberson (University of Georgia) for their

insightful comments on an earlier draft of this manuscript.

References Andow, D. A., Lovei, G. L. & Arpaia, S. (2006). Ecological risk assessment for Bt crops (response to Romeis). Nature Biotechnology, 24, 749–751. Anonymous (1995). Cotton: the crop and its pesticide market. Pesticide News, 30, 1. Baumhover, A. H., Graham, A. J., Bitter, B. A. et al. (1955). Screw-worm control through release of sterilized flies. Journal of Economic Entomology, 48, 462–466. Bottrell, D. G. & Adkisson, P. L. (1977). Cotton insect pest management. Annual Review of Entomology, 22, 451– 481. Brodt, S. J., Goodell, P. B., Krebill-Prather, R. L. & Vargas, R. N. (2007). California cotton growers utilize integrated pest management. California Agriculture, 61, 24– 30. Brown, P., Silvertooth, J., Moore, L. & Watson, T. (1992). Revised planting window for full season cotton varieties. In Cotton: A College of Agriculture Report, pp. 240– 251. Tucson, AZ: University of Arizona Press. Campion, D. G. (1994). Pheromones for the control of cotton pests. In Insect Pests of Cotton, eds. G. A. Matthews & J. P. Tunstall, pp. 505–534. Wallingford, UK: CABI Publishing. Castle, S. J., Prabhaker, N. & Henneberry, T. J. (1999). Insecticide resistance and its management in cotton insects. ICAC Review Article on Cotton Production Research, International Cotton Advisory Committee, 5, 1–55. Christou, P., Capell, T., Kohli, A., Gatehouse, J. A. & Gatehouse, A. M. R. (2006). Recent developments and future prospects in insect pest control in transgenic crops. Trends in Plant Science, 11, 302–308. Conway, H. E., Steinkraus, D. C., Ruberson, J. R. & Kring, T. J. (2006). Experimental treatment threshold for the cotton aphid (Homoptera: Aphididae) using natural enemies in Arkansas cotton. Journal of Entomological Science, 41, 361–373. Cross, W. H. (1973). Biology, control, and eradication of the boll weevil. Annual Review of Entomology, 18, 14–46. Dickerson, W. A., Brashear, A. L., Brumley, J. T. et al. (2001). Boll Weevil Eradication in the United States through 1999, Cotton Foundation Reference Book Series No. 6. Memphis, TN: Cotton Foundation. El-Lissy, O. A. & Grefenstette, W. J. (2006). Progress of pink bollworm eradication in the U.S. and Mexico, 2005, In Proceedings of the Beltwide Cotton Conferences, pp. 1313– 1319. Memphis, TN: National Cotton Council. El-Zik, K. M. & Frisbie, R. E. (1991). Integrated crop management systems for pest control. In Handbook of Pest

COTTON ARTHROPOD IPM Management in Agriculture, vol. 3, ed. D. Pimentel, pp. 3–104. Boca Raton, FL: CRC Press. El-Zik, K. M., Grimes, D. W. & Thaxton, P. M. (1989). Cultural management and pest suppression. In Integrated Pest Management Systems and Cotton Production, eds. R. E. Frisbie, K. M. El-Zik & L. T. Wilson, pp. 11– 36. New York: John Wiley. El-Zik, K. M. & Thaxton, P. M. (1989). Genetic improvement for resistance to pests and stresses in cotton. In Integrated Pest Management Systems and Cotton Production, eds. R. E. Frisbie, K. M. El-Zik & L. T. Wilson, pp. 191–224. New York: John Wiley. Ellsworth, P. C. & Martinez-Carrillo, J. L. (2001). IPM for Bemisia tabaci: a case study from North America. Crop Protection, 20, 853–869. Ellsworth, P. C., Diehl, J. W. & Naranjo, S. E. (1996). Sampling Sweetpotato Whitefly Nymphs in Cotton, IPM Series No. 6, Publication 196006. Tucson, AZ: University of Arizona, Cooperative Extension. Available at http://&break; insects/wfsampl.html. Ellsworth, P. C., Palumbo, J. L., Naranjo, S. E., Dennehy, T. J. & Nichols, R. L. (2006). Whitefly Management in Arizona Cotton 2006, IPM Series No. 18, Publication AZ1404. Tucson, AZ: University of Arizona, Cooperative Extension. Available at insects/wf/ipm6.html. Environmental Protection Agency (2006). Reducing Pesticide Risk. Washington, DC: US Environmental Protection Agency. Available at health/reducing.htm. Eveleens, K. G., van den Bosch, R. & Ehler, L. E. (1973). Secondary outbreak induction of beet armyworm by experimental insecticide application in cotton in California. Environmental Entomology, 2, 497–503. Ewing, K. P. & Parencia, C. R. Jr. (1950). Early-Season Applications of Insecticides on a Community-Wide Basis for Cotton Insect Control in 1950, Bureau of Entomology and Plant Quarantine Publication No. E810. Washington, DC: US Department of Agriculture. Faria, M. & Wraight, S. P. (2001). Biological control of Bemisia tabaci with fungi. Crop Protection, 20, 767– 778. Fillman, D. A. & Sterling, W. L. (1985). Inaction levels for the red imported fire ant, Solenopsis invicta, a predator of the boll weevil, Anthonomus grandis. Agriculture, Ecosystems and Environment, 13, 93–102. Frisbie, R. E., El-Zik, K. M. & Wilson, L. T. (eds.) (1989). Integrated Pest Management Systems and Cotton Production. New York: John Wiley. Frisvold, G. B., Reeves, J. M. & Tronstad, R. (2006). Bt cotton adoption in the United States and China: international trade and welfare effects. AgBioForum, 9, 69–78.

Gaines, J. C. (1957). Cotton insects and their control in the United States. Annual Review of Entomology, 2, 319– 338. Gannaway, J. R. (1994). Breeding for insect resistance. In Insect Pests of Cotton, eds. G. A. Matthews & J. P. Tunstall, pp. 431–453. Wallingford, UK: CABI Publishing. Glare, T. R. & O’Callaghan, M. (2000). Bacillus thuringiensis: Biology, Ecology and Safety. New York: John Wiley. Goodell, P. B., Montez, G. & Wilhoit, L. (2006). Shifting patterns in insecticide use in California, 1993 to 2004. In Proceedings of the Beltwide Cotton Conference, pp. 1367– 1373. Memphis, TN: National Cotton Council. Gould, J., Hoelmer, K. & Goolsby, J. (eds.) (2008). Classical Biological Control of Bemisia tabaci in the United States: A Review of Interagency Research and Implementation. New York: Springer-Verlag. Greene, J. K., Turnipseed, S. G., Sullivan, M. J. & May, O. L. (2001). Treatment thresholds for stink bugs (Hemiptera: Pentatomidae) in cotton. Journal of Economic Entomology, 94, 403–409. Hardee, D. D. & Harris, F. A. (2003). Eradicating the boll weevil (Coleoptera: Curculionidae): a clash between a highly successful insect, good scientific achievement, and differing agricultural philosophies. American Entomologist, 49, 82–111. Hargreaves, H. (1948). List of the Recorded Cotton Insects of the World. London: Commonwealth Institute of Entomology. Harris, F. A., Canerday, T. D., Henry, L. G. & Palmquist, D. L. (1996). Working together: roles of private consultants, industry, researchers, extension, and growers. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 843–851. Memphis, TN: Cotton Foundation. Henneberry, T. J. & Naranjo, S. E. (1998). Integrated management approaches for pink bollworm in the southwestern United States. Integrated Pest Management Review, 3, 31–52. Henneberry, T. J. & Phillips, J. R. (1996). Suppression and management of cotton insect populations on an areawide basis. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 601–624. Memphis, TN: Cotton Foundation. Herzog, G. A., Graves, J. B., Reed, J. T., Scott, W. P. & Watson, T. F. (1996). Chemical control. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 447–469. Memphis, TN: Cotton Foundation. Huber, R. T., Moore, L. & Hoffman, M. P. (1979). Feasibility study of areawide pheromone trapping of male pink




bollworm moths in a cotton insect pest management program. Journal of Economic Entomology, 72, 222–227. Insecticide Resistance Action Committee (2007). MoA Resources. Brussels, Belgium: IRAC. Available at www. Jenkins, J. N. & Wilson, F. D. (1996). Host plant resistance. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 563–600. Memphis, TN: Cotton Foundation. King, E. G., Coleman, R. J., Morales-Ramos, J. A. et al. (1996a). Biological control. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 511–538. Memphis, TN: Cotton Foundation. King, E. G., Phillips, J. R. & Coleman, R. J. (eds.) (1996b). Cotton Insects and Mites: Characterization and Management. Memphis, TN: Cotton Foundation. Knipling, E. F. (1955). Possibilities of insect control or eradication through the use of sexually sterile males. Journal of Economic Entomology, 48, 459–462. Knipling, E. F. (1979). The Basic Principles of Insect Population Suppression and Management, US Department of Agriculture Agricultural Handbook No. 512. Washington, DC: Government Printing Office. Kogan, M. (1998). Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology, 43, 243–270. Kohel, R. J. & Lewis, C. F. (eds.) (1984). Cotton, Monograph No. 24, Agronomy Series. Madison, WI: American Society of Agronomy. Leigh, T. F., Black, H., Jackson, C. E. & Burton, V. E. (1966). Insecticides and beneficial insects in cotton fields. California Agriculture, 20(7), 4–6. Leigh, T. F., Roach, S. H. & Watson, T. F. (1996). Biology and ecology of important insect and mites pests of cotton. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 17–85. Memphis, TN: Cotton Foundation. Luttrell, R. G. (1994). Cotton pest management. II. A US perspective. Annual Review of Entomology, 39, 527–542. Matthews, G. A. (1994). Cultural control. In Insect Pests of Cotton, eds. G. A. Matthews & J. P. Tunstall, pp. 455–461. Wallingford, UK: CABI Publishing. McGuire, M. R., Leland, J. E., Dara, S., Park, Y. H. & Ulloa, M. (2006). Effect of different isolates of Beauveria bassiana on field populations of Lygus hesperus. Biological Control, 38, 390–396. Miller, E., Lowe, A. & Archuleta, S. (2000). Evaluation of different release strategies for use in pink bollworm sterile release programs. In Proceedings of the Beltwide Cotton Conference, pp. 1368–1370. Memphis, TN: National Cotton Council.

Mumford, J. D. & Norton, G. A. (1994). Pest management systems. In Insect Pests of Cotton, eds. G. A. Matthews & J. P. Tunstall, pp. 559–576. Wallingford, UK: CABI Publishing. Myers, J. H., Savoie, A. & van Randen, E. (1998). Eradication and pest management. Annual Review of Entomology, 43, 471–491. Naranjo, S. E. (2001). Conservation and evaluation of natural enemies in IPM systems for Bemisia tabaci. Crop Protection, 20, 835–852. Naranjo, S. E. (2007). Establishment and impact of exotic aphelinid parasitoids in Arizona: a life table approach. Journal of Insect Science, 7, 63 (abstr.). Naranjo, S. E. & Hagler, J. R. (1998). Characterizing and estimating the impact of heteropteran predation. In Predatory Heteroptera: Their Ecology and Use in Biological Control, eds. M. Coll & J. Ruberson, pp. 170–197. Lanham, MD: Entomological Society of America. Naranjo, S. E., Flint, H. M. & Henneberry, T. J. (1996). Binomial sampling plans for estimating and classifying population density of adult Bemisia tabaci on cotton. Entomologia Experimentalis et Applicata, 80, 343–353. Naranjo, S. E., Ellsworth, P. C. & Hagler, J. R. (2004). Conservation of natural enemies in cotton: role of insect growth regulators in management of Bemisia tabaci. Biological Control, 30, 52–72. National Agricultural Statistics Service (2006). Agricultural Chemical Usage: 2005 Field Crops Summary. Washington, DC: US Department of Agriculture. Available at http://usda.mannlib.cornell. edu/usda/nass/&break;AgriChemUsFC/2000s/2006/ AgriChemUsFC-05–17-&break;2006.pdf. National Cotton Council (2007a). Crop Information: Data bases. Memphis, TN: National Cotton Council. Available at National Cotton Council (2007b). Pest Management. Cotton Pest Loss Data. Memphis, TN: National Cotton Council. Available at Oliveira, M. R. V., Henneberry, T. J. & Anderson, P. (2001). History, current status, and collaborative research projects for Bemisia tabaci. Crop Protection, 20, 709–723. Palumbo, J. C., Horowitz, A. R. & Prabhaker, N. (2001). Insecticidal control and resistance management for Bemisia tabaci. Crop Protection, 20, 739–765. Palumbo, J. C., Ellsworth, P. C., Dennehy, T. D. & Nichols, R. L. (2003). Cross-Commodity Guidelines for Neonicotinoid Insecticides in Arizona. IPM Series No. 17, Publication AZ1319. Tucson, AZ: University of Arizona, Cooperative Extension. Available at http://cals. Parvin, D. W. Jr. & Smith, J. W. (1996). Crop phenology and insect management. In Cotton Insects and Mites:


Characterization and Managements, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 815–829. Memphis, TN: Cotton Foundation. Plapp, F. Jr., Jackman, J., Campanhola, C. et al. (1990). Monitoring and management of pyrethroid resistance in the budworm (Lepidoptera: Noctuidae) in Texas, Mississippi, Louisiana, Arkansas and Oklahoma. Journal of Economic Entomology, 83, 335–341. Prasifka, J. R., Heinz, K. M. & Minzenmayer, R. R. (2004). Relationships of landscape, prey and agronomic variables to the abundance of generalist predators in cotton (Gossypium hirsutum) fields. Landscape Ecology, 19, 709–717. Rainwater, C. F. (1952). Progress in research on cotton insects. In Insects, The Yearbook of Agriculture, 1952, pp. 497–500. Washington, DC: US Department of Agriculture, US Government Printing Office. Ridgway, R. L., Bell, A. A., Vetch, J. A. & Chandler, J. M. (1984). Cotton protection practices in the USA and world. In Cotton, Monograph No. 24, Agronomy Series, eds. R. J. Kohel & C. F. Lewis, pp. 266–361. Madison, WI: American Society of Agronomy. Romeis, J., Meissle, M. & Bigler, F. (2006). Transgenic crops expressing Bacillus thuringiensis toxins and biological control. Nature Biotechnology, 24, 63–71. Roush, R. T. & Daly, J. C. (1990). The role of population genetics in resistance research and management. In Pesticide Resistance in Arthropods, eds. R. T. Roush & B. E. Tabashnik, pp. 97–125. New York: Chapman & Hall. Schneider, J. C. (2003). Overwintering of Heliothis virescens (F.) and Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae) in cotton fields of Northeast Mississippi. Journal of Economic Entomology, 96, 1433–1447. Shaw, D. R. & Willers, J. L. (2006). Improving pest management with remote sensing. Outlooks on Pest Management, 17, 197–201. Shelton, A. M. & Badenes-Perez, E. (2006). Concepts and applications of trap cropping in pest management. Annual Review of Entomology, 51, 285–308. Smith, J. W. & Harris, F. A. (1994). Boll weevil: cotton pest of the century. In Insect Pests of Cotton, eds. G. A. Matthews & J. P. Tunstall, pp. 228–258. Wallingford, UK: CABI Publishing. Snodgrass, G. L. (1996). Insecticide resistance in field populations of the tarnished plant bug (Heteroptera: Miridae) in cotton in the Mississippi Delta. Journal of Economic Entomology, 89, 783–790. Snodgrass, G. L., Scott, W. P., Abel, C. A. et al. (2005). Tarnished plant bug (Heteroptera: Miridae) populations near fields after early season herbicide treatment. Environmental Entomology, 34, 705–711.

Sparks, T. C. (1996). Toxicology of insecticides and acaricides. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 283–322. Memphis, TN: Cotton Foundation. Steinkraus, D. C., Hollingsworth, R. G. & Slaymaker, P. H. (1995). Prevalence of Neozygites fresenii (Entomophthorales: Neozygitaceae) on cotton aphids (Homoptera: Aphididae) in Arkansas cotton. Environmental Entomology, 24, 465–474. Sterling, W. L., El-Zik, K. M. & Wilson, L. T. (1989). Biological control of pest populations. In Integrated Pest Management Systems and Cotton Production, eds. R. E. Frisbie, K. M. El-Zik & L. T. Wilson, pp. 155–189. New York: John Wiley. Stern, V., van den Bosch, R. & Leigh, T. F. (1964). Strip cutting alfalfa for lygus bug control. California Agriculture, 18, 5–6. Stoltz, R. L. & Stern, V. M. (1978). Cotton arthropod food chain disruption by pesticides in the San Joaquin Valley, California. Environmental Entomology, 7, 703– 707. Tabashnik, B. E., Carri`ere, Y., Dennehy, T. J. et al. (2003). Insect resistance to transgenic Bt crops: lessons from the laboratory and field. Journal of Economic Entomology, 96, 1031–1038. Toscano, N. C., Mueller, A. J., Sevacherian, V. & Sharma, R. K. (1974). Insecticide applications based on hexalure trap catches versus automatic schedule treatments for pink bollworm moth control. Journal of Economic Entomology, 67, 522–524. Trichilo, P. J. & Wilson, L. T. (1993). An ecosystem analysis of spider mite outbreaks: physiological stimulation or natural enemy suppression. Experimental and Applied Acarology, 17, 291–314. van den Bosch, R. & Hagen, K. S. (1966). Predaceous and Parasitic Arthropods in California Cotton Fields, California Agricultural Experiment Station Bulletin No. 820. Berkeley, CA: University of California. Villavaso, E. J., Bartlett, A. C. & Laster, M. L. (1996). Genetic control. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 539–562. Memphis, TN: Cotton Foundation. Wagner, T. L., Olson, R. L., Willers, J. L. & Williams, M. R. (1996). Modeling and computerized decision aids. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 205–249. Memphis, TN: Cotton Foundation. Walker, J. K. & Smith, C. W. (1996). Cultural control. In Cotton Insects and Mites: Characterization and Management, eds. E. G. King, J. R. Phillips & R. J. Coleman, pp. 471–510. Memphis, TN: Cotton Foundation.




Watson, T. F. (1980). Methods for reducing winter survival of the pink bollworm. In Pink Bollworm Control in the Western United States, ed. H. S. Graham, pp. 24–34. Oakland, CA: US Department of Agriculture, Science and Education Administration, Western Region. Whalon, M. E., Mota-Sanchez, D., Hollingsworth, R. M. & Duynslayer, L. (2007). IRAC Arthropod Pesticide Resistance Database. East Lansing, MI: Michigan State University. Available at Whitcomb, W. H. & Bell, K. (1964). Predaceous Insects, Spiders and Mites of Arkansas Cotton Fields, Bulletin No. 690.

Fayetteville, AR: Arkansas Agricultural Experiment Station. Williams, M. R. (2006). Cotton insect losses 2005. In Proceedings of the Beltwide Cotton Conference, pp. 1151–1204. Memphis, TN: National Cotton Council. Williams, M. R. (2007). Cotton insect losses 2006. In Proceedings of the Beltwide Cotton Conference, pp. 974–1026. Memphis, TN: National Cotton Council. Wilson, L. T., Gonzales, D. & Plant, R. E. (1985). Predicting sampling frequency and economic status of spider mites on cotton. In Proceedings of the Beltwide Cotton Conference, p. 168. Memphis, TN: National Cotton Council.

Chapter 26

Citrus IPM Richard F. Lee IPM programs are designed to keep plants healthy and economically productive while minimizing environmental impact. Citrus (Citrus spp.) as clonally propagated perennial crops are subject to many graft-transmissible diseases caused by viruses, viroids and systemic prokaryotes. Some of these graft-transmissible diseases can be very destructive and even threaten the continued production of citrus in a production area, whereas other diseases cause minor losses. The starting point for an IPM program for citrus is to begin with healthy plants. The concept of planting with healthy plants in citrus began almost simultaneously with the discovery that some diseases of citrus were caused by graft-transmissible pathogens (GTPs). Indicator plants grafted with parts of diseased trees subsequently show characteristic symptoms for the disease (Fawcett, 1938). The concept of a clean stock program evolved from the finding that the disease could be prevented by using graft propagations from source trees that were free of the disease. Applications of this concept led to the development of regional and national clean stock programs and certification programs. GTPs of citrus which do not have insect vectors or other natural means of spread are easily controlled by use of clean, or “pathogentested,” budwood, but diseases which have a natural means of spread, such as insects, are more difficult to control. However, beginning with healthy

plants is even more important for the management and control of these naturally spread grafttransmissible diseases. In this chapter, the essential components of a citrus certification program will be reviewed, the graft-transmissible diseases of citrus and methods available for therapy or cleaning of germplasm will be summarized and the application of mild strain cross protection to lessen losses due to Citrus tristeza virus will be described as an example of the management of a devastating insect vectored graft-transmissible disease.

26.1 Components of a citrus certification program The term “certification program” is often applied in a vague sense. The term has been used to describe a regulatory program to prevent the spread of nematodes affecting citrus, or to recognize the fact that a nursery site has been “certified” to meet certain predefined standards, or to define horticultural standards such as size of propagated plants before sale to growers and height of bud-unions. Here the term, certification program, is used to describe an IPM program designed to produce “pathogen-tested” citrus plants for planting into the field. A typical certification program has three critical elements which are interrelated:

Integrated Pest Management, ed. Edward B. Radcliffe, William D. Hutchison and Rafael E. Cancelado. Published by C Cambridge University Press 2009. Cambridge University Press. 



Exotic germplasm

Quarantine Program

Domestic germplasm

Clean Stock Program

Shoot Tip Grafting/Themotherapy

Select local varieties by defined criteria

Indexing to verify elimination of GTPs

Shoot Tip Grafting/Thermotherapy

Release to Clean Stock Program

Indexing to verify elimination of GTPs Maintain therapied germplasm under protected conditions, re-index at regular intervals Conduct horticultural evaluations Provide nursery material for mother trees in the Citrus Certification Program

Citrus Certification Program Primary Protected Foundation Block Receives nursery material from the Clean Stock Program for establishment of mother trees under protected conditions, regular re-indexing performed for GTPs present in the region, provide budwood/seed for establishment of Foundation Blocks and budwood increase blocks.

Foundation Blocks Originate from budwood and seed from Primary Protected Foundation Block, often owned by private nurserymen but may be maintained by public agencies, usually under protected conditions if insect vectored GTPs are present, regular re-indexing performed for GTPs present in the region, provide budwood for budwood increase blocks and/or certified nursery plants.

Budwood Increase Blocks Provide catalytic increase of budwood obtained from Primary Protected Foundation Block and/or Foundation Block, have a limited lifespan, may be under protected conditions if insect vectored GTPs are present. Provides budwood to propagate Certified Nursery Trees.

Certified Nursery Trees These are healthy citrus trees ultimately planted in the industry, but they may be propagated and grown under protected conditions in areas where insect vectored GTPs are present. Records usually maintained to enable tracing the origin of buds. Fig. 26.1 Diagram of the critical components of a citrus certification program. GTP, graft-transmissible pathogens.

(1) a quarantine program, (2) a clean stock program and (3) a certification program (Lee et al., 1999; Navarro, 1993) (Fig. 26.1).

26.1.1 Quarantine programs Adherence to proper quarantine procedures is an essential safeguard when bringing in new, exotic

germplasm into the local certification program. Citrus growers, by their nature, are always looking for something new and unique, such as lowseeded varieties, varieties higher in color than those presently available, or varieties which may extend the marketing window by maturing earlier and later than varieties presently available. Such germplasm selections will be imported into any citrus industry; it is essential that they come in through an authorized quarantine program so that additional pests and/or graft-transmissible


diseases are not imported with illegally and untested germplasm. Quarantine programs operate under the jurisdiction of the ministry of agriculture of a country or commissioner of agriculture in a state or province. Usually the quarantine program is operated by the plant protection services of a government regulatory agency. The traditional approach to quarantine is to use isolation to prevent accidental entry of new pests into the local industry. The isolation may be geographic with large distances maintained from the industry, or by the use of screened, vector-proof greenhouses. Under quarantine isolation, the new germplasm is tested for presence of pathogens and, if present, then therapy measures are applied to remove the pest(s). Following therapy, retesting is conducted to assure the pest(s) have been eliminated. The best approach for testing is the use of biological indicator plants as their use would enable the visualization of symptoms which may be caused by unsuspected pests. Biological indexing provides a higher level of confidence that all pests have been eliminated. Laboratory testing for pests will reveal freedom or presence only of the pest for which the diagnostic procedure is designed. As an example, a biological index for Citrus psorosis virus on Dweet tangor produces symptoms visually indistinguishable from those caused by several other virus-like pathogens of the psorosis group (e.g. Citrus concave gum, impietratura, cristacortis, and Citrus leaf blotch virus) while the reverse transcription polymerase chain reaction (RT-PCR) assay for psorosis would only detect psorosis. An alternative approach to quarantine by geographic isolation is to use the in vitro method where imported budwood is kept under quarantine in glass test tubes. This procedure was introduced by Navarro in Spain for the safe introduction of new citrus varieties (Navarro et al., 1984; Lee et al., 1999). The budwood, upon arrival, is surface sterilized, then placed in test tubes containing culture media and then maintained in a growth chamber. When the shoots emerge, shoot tip grafting (STG) is done, with the thin section of the meristematic buds being grafted to healthy receptor seedling plants used as a rootstock. This approach offers an advantage in that it requires less space than a quarantine greenhouse, and

entry of new germplasm is usually expedited as everything is shoot tip grafted upon receipt. However, the incoming budwood is not always in good shape following shipment, and sometimes fungal contaminates prevent recovery of the germplasm regardless of efforts to sterilize the budwood stick. Often citrus seed is freely imported without concern to what pests may be present. Several important citrus diseases have been shown to be seed-transmitted including: citrus variegated chlorosis, caused by the xylem-inhabiting bacterium Xylella fastidiosa (Li et al., 2003), witches’ broom disease of lime, caused by Candidatus Phytoplasma aurantifolii (El-Kharbotly et al., 2000; Khan et al., 2002), psorosis and a psorosis-like pathogen in hardy orange (Poncirus trifoliate) and in hybrids having P. trifoliata as one of the parents (e.g. citranges [sweet orange (C. aurantium) × hardy orange] and citrumelos [hardy orange × grapefruit (C. paradisi)] (Roistacher, 1991; Powell et al., 1998), and Citrus leaf blotch virus (Guerri et al., 2004). Care should be taken that seed source trees have been indexed for freedom from these diseases. Additionally imported seed should undergo hot water treatment (10 min at 52 ◦ C), rinse in ambient temperature water, then treatment for 2 min in 1% solution of 8-hydroxquinoline sulfate in water and then air drying to eliminate fungal contamination (Roistacher, 1991).

26.1.2 Clean stock programs Development of the clean stock program has enabled recovery of healthy plants from locally grown (domestic) varieties and cultivars which may be infected with GTPs and the continued maintenance of this therapied germplasm for use in the certification program. Additionally, the clean stock program maintains exotic germplasm which has been previously freed from pathogens during quarantine as “pathogen tested” material for use in the local certification program. Clean stock programs are often maintained by universities, research institutions or non-government organizations as a service for the citrus industry, but they also may be maintained by regulatory agencies. There are several steps involved in the operation of a clean stock program: (1) selection of mother trees from the local cultivars, (2) indexing




of the selected mother trees, (3) therapy to eliminate GTPs which may be present, (4) indexing of the recovered plants to ensure the GTPs have been eliminated, (5) horticultural evaluation of the recovered, healthy plants and (6) maint